Post on 10-Jun-2020
ABCDEFG
UNIVERS ITY OF OULU P.O.B . 7500 F I -90014 UNIVERS ITY OF OULU F INLAND
A C T A U N I V E R S I T A T I S O U L U E N S I S
S E R I E S E D I T O R S
SCIENTIAE RERUM NATURALIUM
HUMANIORA
TECHNICA
MEDICA
SCIENTIAE RERUM SOCIALIUM
SCRIPTA ACADEMICA
OECONOMICA
EDITOR IN CHIEF
PUBLICATIONS EDITOR
Senior Assistant Jorma Arhippainen
Lecturer Santeri Palviainen
Professor Hannu Heusala
Professor Olli Vuolteenaho
Senior Researcher Eila Estola
Director Sinikka Eskelinen
Professor Jari Juga
Professor Olli Vuolteenaho
Publications Editor Kirsti Nurkkala
ISBN 978-951-42-9599-7 (Paperback)ISBN 978-951-42-9600-0 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)
U N I V E R S I TAT I S O U L U E N S I SACTAC
TECHNICA
U N I V E R S I TAT I S O U L U E N S I SACTAC
TECHNICA
OULU 2011
C 396
Timo Koskela
COMMUNITY-CENTRIC MOBILE PEER-TO-PEER SERVICES: PERFORMANCE EVALUATION AND USER STUDIES
UNIVERSITY OF OULU,FACULTY OF TECHNOLOGY,DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING;INFOTECH OULU
C 396
ACTA
Tim
o Koskela
C396etukansi.kesken.fm Page 1 Monday, October 17, 2011 11:53 AM
A C T A U N I V E R S I T A T I S O U L U E N S I SC Te c h n i c a 3 9 6
TIMO KOSKELA
COMMUNITY-CENTRIC MOBILE PEER-TO-PEER SERVICES: PERFORMANCE EVALUATIONAND USER STUDIES
Academic dissertation to be presented with the assent ofthe Faculty of Technology of the University of Oulu forpublic defence in Auditorium IT116, Linnanmaa, on 23November 2011, at 12 noon
UNIVERSITY OF OULU, OULU 2011
Copyright © 2011Acta Univ. Oul. C 396, 2011
Supervised byDocent Mika Ylianttila
Reviewed byDoctor Peter ReichlProfessor Kaisa Väänänen-Vainio-Mattila
ISBN 978-951-42-9599-7 (Paperback)ISBN 978-951-42-9600-0 (PDF)
ISSN 0355-3213 (Printed)ISSN 1796-2226 (Online)
Cover DesignRaimo Ahonen
JUVENES PRINTTAMPERE 2011
Koskela, Timo, Community-centric mobile peer-to-peer services: performanceevaluation and user studies. University of Oulu, Faculty of Technology, Department of Computer Science and Engineering;Infotech Oulu, P.O. Box 4500, FI-90014 University of Oulu, FinlandActa Univ. Oul. C 396, 2011Oulu, Finland
AbstractThe popularity of social networking services emphasises the importance of user communities. Asthe next evolutionary step, social networking will evolve towards user communities that aredynamically established according to contextual aspects such as the shared location and/oractivities of people. In this vision, distributed peer-to-peer (P2P) technologies provide arespectable alternative to the centralised client-server technologies for implementing theupcoming community-centric mobile services.
This thesis focuses on the development of enabling techniques and methods for a community-centric mobile service environment and on the evaluation of users’ perceptions and useracceptance of novel community-centric mobile services. The research was carried out by firstexamining the applicability of different P2P system architectures for implementing themanagement of user communities. Next, an experimental prototype of a community-centricmobile service environment was implemented. The experimental prototype was used forevaluating the performance of the community-centric mobile service environment and forconducting the user studies of novel community-centric mobile services. The evaluation wasconducted in a real deployment environment with test users.
The main results are the following: (1) a conceptual analysis of the advantages anddisadvantages of P2P group management systems that are implemented using architecturallydistinct P2P system architectures; (2) a performance evaluation of a P2P community managementsystem, in which each community is implemented as an independent structured P2P overlaynetwork; (3) a novel community-centric mobile service environment and its core component, amobile middleware, which enables the use of P2P technologies and context information as part ofWeb-based applications running on a Web browser; (4) an evaluation of users’ perceptions anduser acceptance of novel community-centric mobile services that are related to making a selectionof an entertainment premise based on its music style and voting for music in a user community.
Keywords: community management, context information, mobile middleware, mobileservices, peer-to-peer networks, user studies, user-centred design, web 2.0
Koskela, Timo, Yhteisölliset mobiilit vertaisverkkopalvelut: suorituskyvynarviointi ja käyttäjätutkimus. Oulun yliopisto, Teknillinen tiedekunta, Tietotekniikan osasto; Infotech Oulu, PL 4500, 90014Oulun yliopistoActa Univ. Oul. C 396, 2011Oulu
TiivistelmäKäyttäjäyhteisöiden tärkeys on korostunut sosiaalista verkottumista edistävien palveluiden suo-sion kautta. Seuraavaksi sosiaalinen verkottuminen tulee laajentumaan yhteisöihin, jotka perus-tetaan dynaamisesti muun muassa ihmisten yhteisen sijainnin ja/tai toiminnan perusteella. Tule-vaisuuden yhteisöllisten mobiilipalveluiden toteutuksessa hajautetut vertaisverkkoteknologiattarjoavat varteenotettavan vaihtoehdon keskitetyille asiakas-palvelin -teknologioille.
Tämä väitöskirjatyö keskittyy kehittämään tarvittavia tekniikoita ja menetelmiä yhteisöllisenmobiilipalveluympäristön toteuttamiseen sekä arvioimaan käyttäjien kokemuksia uusista yhtei-söllisistä mobiilipalveluista. Tutkimuksessa tarkasteltiin aluksi arkkitehtuurillisesti erilaistenvertaisverkkojärjestelmien sopivuutta käyttäjäyhteisöjen hallinnan toteutettamiseksi. Seuraavak-si tutkimuksessa toteutettiin kokeellinen prototyyppi yhteisöllisestä mobiilipalveluympäristöstä.Kokeellista prototyyppiä hyödynnettiin sekä yhteisöllisen mobiilipalveluympäristön suoritusky-vyn että uusien yhteisöllisten mobiilipalveluiden käyttäjäkokemuksen arvioinnissa. Arviointisuoritettiin aidossa käyttöympäristössä testikäyttäjien avulla.
Väitöskirjatyön keskisimmät tulokset ovat seuraavat: (1) käsitteellinen analyysi arkkitehtuu-rillisesti erilaisten vertaisverkkojärjestelmien vahvuuksista ja heikkouksista ryhmänhallintajär-jestelmän toteuttamisessa; (2) sellaisen yhteisönhallintajärjestelmän arviointi, jossa jokainenyhteisö on toteutettu erillisen ja itsenäisen rakenteellisen vertaisverkon avulla; (3) uusi yhteisöl-linen mobiilipalveluympäristö, jonka keskeisin osa mobiili välikerrosohjelmisto mahdollistaavertaisverkkoteknologioiden ja kontekstitiedon käyttämisen selaimessa toimivissa Web-sovel-luksissa; (4) sellaisten uusien yhteisöllisten mobiilipalveluiden käyttäjäkokemuksen arviointi,jotka liittyvät vapaa-ajan viettopaikan valitaan musiikin perusteella ja soitettavan musiikinäänestämiseen käyttäjäyhteisöissä.
Asiasanat: kontekstitieto, käyttäjälähtöinen suunnittelu, käyttäjätutkimukset,mobiilipalvelut, mobiilit välikerrosohjelmistot, verkkoyhteisöt, vertaisverkot, web 2.0
7
Acknowledgements
The research work presented in this thesis has been carried out at the MediaTeam
Oulu research group, Department of Computer Science and Engineering,
University of Oulu, Finland. Specifically, the research work was conducted as
part of Eureka ITEA2 Expeshare and DECICOM projects.
First, I would like to express my gratitude to my supervisor, Doc. Mika
Ylianttila who has inspired me to achieve my best and provided me valuable
guidance throughout my post-graduate studies. In addition, Mrs. Sari Järvinen, Mr.
Jari Korhonen, Ms. Nonna Kostamo, Ms. Meirong Liu, Mr. Juuso Ohtonen, Dr.
Zhonghong Ou, Mr. Jani Pellikka, and especially Mr. Erkki Harjula, Mr. Janne
Julkunen and Mr. Otso Kassinen, deserve my appreciation for their invaluable
contributions to the original publications in this thesis. I am also grateful to the
research and industrial partners in Expeshare project, and especially to Lic. Johan
Plomp from VTT, Mr. Panu Kanervo from Nokia, Mr. Pablo Harju from Kutalab
for successful collaboration.
I would like to thank Professor Antti Ylä-Jääski for serving as the opponent in
the doctoral defence. I would also like to thank the official reviewers, Dr. Peter
Reichl from Telecommunications Research Center Vienna and Professor Kaisa
Väänänen-Vainio-Mattila from Tampere University of Technology for their
valuable feedback.
For supporting this thesis financially, I am very grateful to Graduate School
in Electronics, Telecommunications and Automation (GETA), Infotech Oulu
Graduate School, Walter Ahlström Foundation, Tauno Tönning Research
Foundation, Finnish Foundation for Technology Promotion, Nokia Foundation,
Research and Training Foundation of TeliaSonera, HPY Research Foundation,
Riitta and Jorma J. Takanen Foundation, Finnish Funding Agency for Technology
and Innovation (Tekes), and the industrial partners in DECICOM project,
including Ericsson, Nokia and Nethawk.
I would like to deeply thank my parents Aila and Erkki, my parents-in-law
Ailakki and Kari, as well as a family friend Dr. Jukka Kortelainen for all their
support throughout my post-graduate studies. Finally, my wife Saija deserves my
most sincere appreciation for her encouragement and patience especially in the
times of hardship. Without you this would not have been possible.
Oulu, October 2011 Timo Koskela
8
9
Abbreviations
AJAX Asynchronous JavaScript and XML
API Application Programming Interface
BT Bluetooth
CCMV Community-centric Music Voting
CNL City Night Life
COMSE Community-centric Mobile Service Environment
CS Client-Server
DHT Distributed Hash Table
DTPB Decomposed Theory of Planned Behaviour
ESB Enterprise Service Bus
FW Firewall
GMS Group Management System
GPS Global Positioning System
HCI Human-Computer Interaction
HTML Hypertext Markup Language
HTTP Hypertext Transfer Protocol
ICT Information and Communication Technology
IDT Innovation Diffusion Theory
IEEE Institute of Electrical and Electronics Engineers
IP Internet Protocol
IS Information System
ISO International Organization for Standardization
ITU International Telecommunication Union
J2ME Java Platform, Micro Edition
MANET Mobile Ad hoc Network
MIS Mobile Information System
NAT Network Address Translation
ORM Object-Role Modelling
OS Operating System
P2P Peer-to-Peer
PCMS P2P Community Management System
PIDP Presence Information Data Format
PIDP-LO Presence Information Data Format Location Object
RFID Radio Frequency Identification
QoE Quality of Experience
10
QoS Quality of Service
SIMPLE Session Initiation Protocol for Instant Messaging and Presence
Leveraging Extensions
SIP Session Initiation Protocol
SNS Social Networking Service
SOA Service-Oriented Architecture
TAM Technology Acceptance Model
TPB Theory of Planned Behaviour
TRA Theory of Reasoned Action
TTL Time-To-Live
UCD User-Centred Design
UDDI Universal Description, Discovery and Integration
UIA Unmanaged Internet Architecture
UML Unified Modelling Language
URL Uniform Resource Locator
UTAUT Unified Theory of Acceptance and Use of Technology
UX User Experience
WAP Wireless Application Protocol
WSDL Web Service Description Language
WLAN Wireless Local Area Network
XML Extensible Markup Language
11
List of original publications
This thesis is based on the following original publications, which are referred to
in the text by Roman numerals (I-VII):
I Koskela T, Kassinen O, Harjula E & Ylianttila M (In press) P2P Group Management Systems: A Conceptual Analysis. ACM Computing Surveys.
II Koskela T, Kassinen O, Ou Z & Ylianttila M (2010) Improving Community Management Performance with Two-level Hierarchical DHT Overlays. Journal of Internet Technology, Special Issue on Internet Resource Sharing and Discovery 11(2): 167–179.
III Koskela T, Harjula E, Kassinen O & Ylianttila M (2011) Robustness of a P2P Community Management System based on Two-level Hierarchical DHT Overlays. Proc IEEE Symposium on Computers and Communication. Kerkyra, Greece: 881–886.
IV Koskela T, Kostamo N, Kassinen O, Ohtonen J & Ylianttila M (2007) Towards Context-Aware Mobile Web 2.0 Service Architecture. Proc IARIA International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies. Papeete, French Polynesia: 41–48.1
V Koskela T, Kassinen O, Harjula E, Pellikka J & Ylianttila M (2010) Case Study on a Community-Centric Mobile Service Environment. International Journal of Web Applications 2(3): 187–205.
VI Koskela T, Julkunen J, Kassinen O & Ylianttila M (2010) User Evaluation of a Community-centric Music Voting Service. Journal of Digital Content Technology and its Application 4(1): 69–78.
VII Koskela T, Järvinen S, Liu M & Ylianttila M (2010) User Experience in Added Value Location-based Mobile Music Service. Proc IEEE International Conference on Web Services. Miami, USA: 465–472.
Papers I-III focus on the research challenges related to P2P group management.
Papers IV and V address the research issues on the development of a service
environment for community-centric mobile services. Papers VI and VII examine
the factors that affect user acceptance of community-centric mobile services.
1 The best paper award
12
13
Table of contents
Abstract
Tiivistelmä
Acknowledgements 7 Abbreviations 9 List of original publications 11 Table of contents 13 1 Introduction 15
1.1 Background ............................................................................................. 15 1.2 Motivation and research problems .......................................................... 17 1.3 Scope and methodology .......................................................................... 19 1.4 Contributions of the thesis ...................................................................... 21 1.5 Organisation of the thesis ........................................................................ 24
2 Literature review 25 2.1 Group management in mobile environments .......................................... 25
2.1.1 Multi-disciplinary nature of groups .............................................. 25 2.1.2 Characteristics of a group management system ............................ 28 2.1.3 Technologies for group management systems .............................. 29
2.2 Convergence of technologies in mobile environments ........................... 36 2.2.1 Benefits of mobile middleware .................................................... 36 2.2.2 Versatile role of context information ............................................ 38 2.2.3 Integrated use of P2P and Web technologies ................................ 42
2.3 Evaluating user acceptance of information systems ................................ 44 2.3.1 Defining user experience .............................................................. 44 2.3.2 User-centred design ...................................................................... 45 2.3.3 Models of user acceptance for information systems ..................... 46
3 Summary of the research contributions 53 3.1 Conceptual analysis of P2P group management systems ........................ 53 3.2 Performance evaluation of a P2P community management
system based on two-level hierarchical DHT overlays ........................... 54 3.3 A novel community-centric mobile service environment ....................... 57 3.4 User acceptance studies of community-centric mobile services ............. 60 3.5 Discussion and future work ..................................................................... 64
4 Conclusions 67 References 69 Original publications 79
14
15
1 Introduction
1.1 Background
“There is no reason anyone would want a computer in their home”
Ken Olson 1977
Only two years after Ken Olson’s statement, the first commercially automated
cellular network was launched in Tokyo, Japan. Three decades later, mobile
phones had already reached the global market penetration rate of 50%. Along
with the huge success at the global market, the capabilities of mobile devices have
evolved exceptionally fast. Today, modern mobile devices are capable of sensing
their environment and connecting to the Internet using multiple wireless network
interfaces. In addition, screen size and resolution as well as input mechanisms
have advanced significantly during the recent years. As a consequence, we not
only have a computer at home, but also in our very own pocket.
The vast popularity of mobile devices has opened a completely new
playground for both research and business people who have struggled to develop
novel mobile services to answer to customers’ needs and to create new ones. The
growth expectations for the mobile service market were at their highest at the turn
of the 21st century, right before the burst of the information and communication
technology (ICT) bubble. As a consequence, the scepticism towards the mobile
services increased for several years to come (Yuan & Zhang 2003).
What makes mobile service market especially challenging is the novelty of
the services themselves. This is because of the fact that many mobile services are
completely new to the users, and thus, users’ perceptions on their added value are
hard to predict (Kesti et al. 2004, Fouskas et al. 2005). The growth of the mobile
service market has also been interfered by the difficulty of finding and installing
new mobile applications (Kassinen et al. 2005) as well as the slow penetration of
3G-capable mobile devices (Tallberg et al. 2007). These two issues are somewhat
intertwined, because without interesting mobile services there is no incentive to
invest in an expensive 3G-capable mobile device. The penetration of 3G-capable
mobile devices has been especially slow in countries, where bundling of mobile
subscriptions with mobile devices was still quite recently prohibited by law
(Tallberg et al. 2007).
The availability of mobile services has been significantly improved by mobile
device manufacturers and mobile operating system (OS) providers (e.g. Nokia,
16
Apple and Google) who have established their own virtual market places, where
mobile applications can be easily acquired. Today, not only the official service
providers, but also any capable user can create their own mobile applications and
gain profit by selling them effortlessly in the virtual market place. This opening
has significantly increased the number of available mobile applications and has
had a positive effect on the development of the mobile service market at a global
scale. The popularity of creating home-made mobile applications has also been
supported by versatile development tools that have been made available for
everyone, and most importantly, mainly free of charge.
In 2003, while the mobile service market was still suffering from its growing
pains, the social networking services (SNSs) started to emerge in the Internet. By
definition, an SNS is a Web-based service that maintains a list of users, their
personal profiles and their connections with each other (Boyd & Ellison 2007). In
the late 2006, social networking was revolutionised when an SNS called
Facebook was opened for everyone. Its success has been tremendous, because
today Facebook has more than 500 million active users and is estimated to be
worth of 50 billion dollars (Facebook 2011).
Facebook is one of the many examples that highlight the importance of
communities in today’s online services. Although any SNS can also be used with
high-end mobile devices, few SNSs take advantage of the information that is
produced by a mobile user. The information that characterises the situation of an
entity such as a user is called context information (Dey et al. 2001).
It is predicted that the social networking phenomenon will eventually expand
to mobile environments, which opens new possibilities for the growth of the
mobile service market. Use of context information as a building block in an SNS
enables the implementation of new kinds of community-centric mobile services
that are established according to the shared location and/or activities of people
(Tasch & Brakel 2004, Morr & Kawash 2007, Nguyen et al. 2008). In this context,
the term community-centric refers to a commercial or a non-profit service that is
built around a user community whose members provide added-value to the
service itself. The added-value may result from the mere participation in the
service usage or from the provided user-specific information and/or user-created
content. The future implications suggest that these communities become an
invisible and integrated part of our physical environment (Nguyen et al. 2008).
Although the most of the existing SNSs rely on a client-server (CS) based
communication model, the feasibility of using decentralised communication
models based on either peer-to-peer (P2P) (Androutsellis-Theotokis & Spinellis
17
2004) or mobile ad hoc network (MANET) (Chlamtac et al. 2003) technologies
should be further examined. One of the main benefits of using P2P or MANET
technologies in community-centric mobile services is that they are low cost in
terms of required hardware and its maintenance (Leibnitz et al. 2007).
The multifaceted research field of this thesis is examined from the viewpoint
of computer networks engineering, but also some elements from mobile
information systems (MIS) (Krogstie et al. 2003, Cheong & Park 2005, Fouskas
et al. 2005) and human-computer interaction (HCI) (Dix et al. 2004, Forlizzi &
Battarbee 2004) research are adopted. When developing novel community-centric
mobile services, the impact of technological choices, the factors affecting user
acceptance and the viability of potential business models should not be
considered separately, but as an integrated task.
1.2 Motivation and research problems
Today, every fifth mobile phone sold is a smartphone that is a handheld computer
embedded with the functionality of a mobile phone. Typically, a smartphone is
equipped with multiple sensors as well as wireless network interfaces, and is
capable of running a large variety of mobile applications. Smartphones are
quickly becoming the norm, because it is expected that the smartphone market
will grow four times faster than the market for traditional mobile handsets
(International Data Corporation 2011). However, the smartphone market is still
populated by multiple competing OSs, which complicates the development of
interoperable mobile services. As a consequence, this trend provides incentive for
using Web technologies for developing mobile services, because Web-based
services are inherently platform-independent.
Fuelled by the growing number of smartphones with built-in GPS devices,
location-based services are transforming from the research concepts into reality
(Junglas & Watson 2008). Despite the recent development, location information
can still be used in a more versatile manner in activity- and context-aware
applications, for instance, to enhance social relationships (Dey et al. 2010). Based
on this trend and the huge success of SNSs, the social networking phenomenon is
predicted to expand to communities that are established according to the shared
location and/or activities of people (Tasch & Brakel 2004, Morr & Kawash 2007,
Nguyen et al. 2008).
To facilitate the development of the emerging community-centric mobile
services, research on the service environment, and especially on the alternative
18
communication models, is needed. Distributed P2P technologies provide a strong
alternative to the centralised CS technologies for managing user communities and
enabling interaction between the community members. According to
Androutsellis-Theotokis & Spinellis (2004), P2P technologies are, inherently,
viable platforms for both distributed computing and interpersonal communication.
Structured P2P networks, in which each network resource has its globally unique
position in the numeric namespace, enable deterministic resolution of search
queries. This guarantees that all groups, group members and the provided services
can be reliably found in the community-centric service environment. Structured
P2P networks, however, can inflict a relatively large amount of load on mobile
devices due to the globally maintained data structure. This is a concern that needs
to be addressed when developing a community management system using
structured P2P technologies.
From the business standpoint, the most important advantages of using P2P
technologies in community-centric mobile services are low hardware and
maintenance costs (Leibnitz et al. 2007). This not only encourages, but also
speeds up the introduction of new community-centric mobile services to the
market. Based on this and the above mentioned remarks, the following research
questions are addressed in this thesis:
1. How to efficiently establish communities using different types of P2P system
architectures?
2. How to implement a robust and efficient community management system for
mobile nodes using structured P2P technologies?
3. What kind of communication infrastructure is required for implementing a
community-centric mobile service environment using P2P and Web
technologies?
4. What kind of factors affect user acceptance of exemplary community-centric
mobile services?
In Fig. 1, the original publications are mapped to the research areas and the
research questions of this thesis. P2P group management and the integrated use of
P2P and Web technologies provide technical enablers for the community-centric
mobile service environment (COMSE). Exemplary community-centric mobile
services, in turn, provide an instrument for evaluating both the feasibility of the
COMSE and the user acceptance of this type of new mobile services.
19
Fig. 1. The mapping of the original publications to the research areas and the
research questions of this thesis.
1.3 Scope and methodology
This thesis focuses on developing a service environment for community-centric
mobile services and examining the factors that affect user acceptance of novel
community-centric mobile services. The communication model for community-
centric mobile services can be implemented using CS, MANET or P2P
technologies, whereas the user interface can be implemented either as a local or a
Web-based application. In this thesis, the focus is especially on the integrated use
of P2P and Web technologies. Following the principles of constructive research,
the described research topics are primarily examined using real-life prototype
implementations. In order to answer to the research questions specified in Section
1.2, three different methodological approaches are used.
First, conceptual analysis is used in Paper I for examining how group
management can be efficiently implemented using different kinds of P2P system
architectures. The analysis is conducted using performance, robustness, fairness,
suitability for battery-powered devices, scalability and security as evaluation
criteria. The selected criteria present the essential challenges faced in P2P
networking (Androutsellis-Theotokis & Spinellis 2004, Kelényi & Nurminen
Paper I
RQ1
Paper II
RQ1, RQ2
Paper III
RQ2
RQ3
Paper VPaper IV
RQ3
RQ4
Paper VI
RQ4
Paper VII
P2P group managementCommunity-centric
mobile services
Community-centric mobile service environment
Integrated use ofP2P and Web technologies
20
2008, Leonard et al. 2008). In addition, conceptual analysis is also used in Paper
IV for examining what kinds of market structures are likely to appear when
different communication models are used for community-centric mobile services.
The market structures were analysed using the theory of industrial organisation.
However, this analysis was not done by the author, and thus, is not discussed in
detail.
Second, real-life prototypes of a P2P community management system (PCMS)
and the COMSE were implemented to be able to analyse their performance,
robustness, and energy-efficiency in Papers II, III and V. It must be noted that the
used distributed hash table (DHT)-based P2P protocol called P2PP (Baset et al.
2007) was not developed by the author, but was used as an essential building
block in both prototype implementations. The used DHT algorithms were
Kademlia (Maymounkov & Mazieres 2002) and Chord (Stoica et al. 2001). The
prototype of the PCMS was evaluated by running multiple instances of P2P nodes
within a dedicated server machine, whereas the prototype of the COMSE was
evaluated in an authentic deployment environment with real test users. The
mobile devices used by the test users were connected to the Internet using a
wireless local area network (WLAN). The established P2P overlay networks were
of different size and contained varying levels of node activity. The prototypes
were evaluated using metrics common to the research community of P2P
networking, namely network traffic load, routing latency, look-up hop count, and
success ratio of look-up operations (Xu et al. 2003, Ou et al. 2010).
Third, user studies and two real-life prototype services are used in Papers VI
and VII for studying users’ perceptions and the factors that affect user acceptance
of exemplary community-centric mobile services. The exemplary services,
Community-centric Music Voting (CCMV) and City Night Life (CNL), were
developed by following some of the key principles of the user-centred design
(UCD) approach. In the development process, a scenario-based description of the
services was first defined in collaboration with the industrial partners relying on
their knowledge on the potential customers. The development of the services was
conducted incrementally, and the testing of the services was implemented in two
iterations. The preliminary testing of the CCMV service was conducted in a
laboratory environment with eleven test users (Koskela et al. 2009), and the final
testing of both CCMV and CNL services, in a real deployment environment with
53 test users. The final testing was conducted in five consecutive testing events,
of which each was populated by approximately ten test users. Both prototype
services were used during each testing event by every test user. First, a test user
21
used the CNL service to find a bar matching her music taste, and second, the
CCMV service to interact in a user community of a pre-selected bar and vote for
music to be played in that bar. Each testing event lasted approximately 45 minutes.
Although the duration of a testing event was fairly short, for instance, Koivumäki
et al. (2008) point out that users’ perceptions on a mobile service do not notably
depend on the length of the test period. This clearly emphasises the importance of
the first impression.
The research data was collected in two phases: first, background information
about test users’ technological awareness and music consumption habits was
collected beforehand using an online questionnaire. Second, users’ perceptions on
the services were collected after the testing event using a quantitative
questionnaire with a 5-point Likert-scale, ranging from strongly disagree to
strongly agree. Finally, short group interviews were conducted that focused on
users’ perceptions on the services, but also on the ideas for future improvements.
For analysing the research data concerning the CCMV service, a research
model called Decomposed Theory of Planned Behaviour (DTPB) (Taylor & Todd
1995) was selected. DTPB examines the effects of attitude, social influence and
perceived behavioural control on the usage intentions towards the service under
investigation. The selection of the research model was based on the facts that
DTPB: (1) performs well in studies focusing on customer behaviour, (2) suits well
for studies, where other individuals may have an influence on the use behaviour
of MISs, (3) can also be applied when the studied MIS is still in the development
phase, and (4) is not overly complex, and thus, rather straight-forward to
implement (Teo & Pok 2003). In addition, the research data concerning both the
CCMV and the CNL services was examined by defining special interest groups
based on test users’ technological awareness and music consumption habits. This
was done, because the special interest groups were predicted to have different
kinds of expectations and attitudes towards the two exemplary community-centric
mobile services. Furthermore, the aim was to find out the user groups that were
likely to be the early adopters of the CCMV and the CNL services.
1.4 Contributions of the thesis
The main contributions are summarised in this section. Specific contributions of
each original paper are elaborated in more detail in Chapter 3. The main
contributions of this thesis are the following:
22
1. Conceptual analysis of P2P group management systems
2. Performance evaluation of a P2P community management system based on
two-level hierarchical DHT overlays
3. A novel community-centric mobile service environment
4. User acceptance studies on exemplary community-centric mobile services.
Paper I introduces a conceptual analysis of P2P group management systems
(GMSs). The publication illustrates how groups are formed using different P2P
system architectures, and analyses the advantages and disadvantages of using
each P2P system architecture in the implementation of a P2P GMS. The
guidelines given in the publication facilitate the selection of an appropriate P2P
system architecture for implementing a P2P GMS. The author was responsible for
conducting the conceptual analysis of P2P GMSs. Mr. Kassinen and Mr. Harjula
defined the evaluation criteria used in the conceptual analysis and participated in
the finalisation of the publication. Doc. Ylianttila was the supervisor.
Paper II introduces and evaluates the PCMS that is a system for managing
P2P communities. In the PCMS, a separate DHT overlay is established for each
community, resulting in multiple small overlays that are indexed and discovered
using a global DHT overlay. The PCMS was evaluated using P2PP as the DHT-
based P2P protocol and Kademlia as the DHT algorithm. The performance of the
PCMS was measured and evaluated in terms of network traffic load, routing
latencies, success ratio of look-up operations and mobile nodes’ energy
consumption. In the evaluation, the PCMS is also compared against a flat DHT
structure, where all the community related activities are performed in a single
DHT overlay. The author was responsible for designing the architecture, defining
the test setup and analysing the evaluation results. Mr. Korhonen and Mr.
Kassinen implemented the test setup and performed the test cases. Dr. Ou was
responsible for the latency analysis. Doc. Ylianttila was the supervisor and
participated in the finalisation of the publication with Mr. Kassinen.
Paper III continues the work started in Paper II and evaluates the robustness
of the PCMS more thoroughly. The evaluation is based on the measurements
conducted using P2PP with two DHT algorithms called Kademlia and Chord. The
robustness of the PCMS is evaluated with different levels of churn (i.e. constant
joining and leaving of nodes) and node activity, in terms of nodes’ look-up
success ratio, network traffic load and average hop count. The PCMS is also
compared against a flat DHT structure where all the community related activities
are performed in a single DHT overlay. The author was responsible for defining
23
the test setup and analysing the evaluation results. Mr. Julkunen implemented the
test setup, whereas Mr. Harjula and Mr. Kassinen participated in the research
discussion and the finalisation of the publication. Doc. Ylianttila was the
supervisor.
Paper IV describes a context-aware mobile Web 2.0 service architecture that
connects individual and community context information with Web-based services.
Based on the degree of centralisation, four communication models for indexing
and discovery of context information and Web-based services are introduced. In
addition, market structures of the proposed communication models are analysed
from a multidisciplinary point of view. The author was responsible for designing
the architecture and the four communication models, whereas Ms. Kostamo
conducted the market structure analysis. Mr. Kassinen and Mr. Ohtonen
participated in the research discussion and in the finalisation of the publication.
Doc. Ylianttila was the supervisor.
Paper V continues the work started in Paper IV and introduces the COMSE, a
service environment for community-centric mobile services that are developed
with Web technologies. The COMSE was built on a distributed communication
infrastructure implemented using P2PP and Kademlia. The interaction between
P2P overlay networks and Web-based services is realised with a mobile
middleware component called P2P Daemon. The performance of the COMSE was
evaluated in terms of network traffic load, success ratio of P2P operations, and
latencies of successful P2P operations. The author was responsible for designing
the COMSE, defining the test setup and analysing the evaluation results. Mr.
Julkunen, Mr. Kassinen and Mr. Pellikka implemented a prototype of the COMSE,
whereas Mr. Harjula participated in the finalisation of the publication. Doc.
Ylianttila was the supervisor.
Paper VI examines the factors that affect user acceptance of the CCMV
service that relies on the COMSE and a virtual disc jockey system called
DjOnline. The CCMV service enables members of a community to influence the
music played within their premises and to interact with each other. The testing of
the CCMV service was conducted in a real usage environment, where the user
communities were comprised of the clientele of a bar. In the evaluation, a user
acceptance model called DTPB was used. The author was responsible for
gathering the research data and studying the factors that affect user acceptance of
the CCMV service. Mr. Julkunen implemented a prototype of the CCMV service,
whereas Mr. Kassinen participated in the finalisation of the publication. Doc.
Ylianttila was the supervisor.
24
Paper VII studies users’ perceptions on the CNL service that relies on the
COMSE and DjOnline. The CNL service provides recommendations about
entertainment premises in the proximity by matching their music styles with the
user’s personal music preferences. Entertainment premises are shown on a map
interface and are marked with colours according to their match. The testing of the
CNL service was conducted in a real usage environment. The analysis of users’
perceptions was based on different user groups that were established according to
test users’ technological awareness and music consumption habits. The author
was responsible for designing the CNL service, gathering the research data and
conducting the analysis of users’ perceptions on the CNL service. Mrs. Sari
Järvinen conducted a user study on the usage of personal music profiles, whereas
Ms. Liu participated in the development of the matching algorithm. Doc.
Ylianttila was the supervisor and participated in the finalisation of the publication.
1.5 Organisation of the thesis
The thesis is organised as follows: in this chapter, the background of the research
topics, motivation and research problems, scope and methodology, as well as a
brief summary of the contributions of this thesis are discussed.
In Chapter 2, a literature review of the research topics covered in this thesis is
presented. The research topics include group management in mobile
environments, convergence of technologies in mobile environments and
evaluating user acceptance of information systems.
In Chapter 3, the main research contributions of the original publications are
summarised, the significance of the results is discussed and ideas for future
research are presented. This chapter consists of five parts that are conceptual
analysis of P2P group management systems, performance evaluation of a P2P
community management system based on two-level hierarchical DHT overlays, a
novel community-centric mobile service environment, user acceptance studies of
community-centric mobile services as well as discussion and future work.
In Chapter 4, some conclusions are finally drawn based on the results of the
thesis and the current development of the community-centric mobile services.
25
2 Literature review
This chapter provides an overview of the literature related to the community-
centric mobile services. The chapter is divided into three parts: Section 2.1
examines the group management in mobile environments, Section 2.2 investigates
the convergence of technologies in mobile environments and Section 2.3 deals
with user studies, having the main focus on the different models for assessing user
acceptance of information technology.
2.1 Group management in mobile environments
“A community is more than the sum of its parts”
Paul Lample 2000
Although the concept of a group may seem simple and well-defined, it is actually
rather ambiguous. This is because of the fact that the concept of a group has
diverse meanings for professionals of different sciences. In this section, the
concepts of a group and a community are defined from a multi-disciplinary
viewpoint and the basic characteristics of group management systems are
presented together with three fundamental enabling technologies.
2.1.1 Multi-disciplinary nature of groups
The concept of a group has particular meanings for professionals of different
sciences (Gupta & Kim 2004). For a psychologist, a group usually refers to two
or more people. For an engineer, in turn, a group may refer to two or more users
or computing entities, whereas an economist primarily interprets a group as two
or more customers whose demographics or consumption behaviour are alike.
In a similar manner, misapprehensions also take place between the concepts
of a group and a community. The two concepts are often used synonymously,
although they have a different meaning (Khambatti et al. 2002). Whereas a group
consists of two or more people or computing entities that may be unknown or at
least anonymous to each other (Tasch & Brakel 2004), a community consists of
two or more people who interact around a shared interest according to some
agreed rules of conduct (Porter 2004). Based on the two definitions, the concept
of a community can be considered as a special case of the concept of a group.
Consequently, all communities are groups, but not all groups are communities.
26
The growing popularity of personal computers and the Internet in the 1990’s
enabled establishing communities also in the electronic frontier (Rheingold 2000).
As a consequence, Porter (2004) categorises communities into physical, partially
virtual and virtual counterparts depending on the used medium for
communication. A community can be defined as (completely) virtual when all
communication is mediated by technology. This varying degree of virtualisation is
also introduced by Morr & Kawash (2007) in their classification of communities.
It is also important to take into account the mobility of the used communication
technology. Thus, technology as a communication medium can be further divided
into mobile and stationary devices (Tasch & Brakel 2004). The use of a mobile
device enables, for instance, more spontaneous communication patterns and
implementation of services that are related to specific locations.
Another distinguishing characteristic of virtual communities is the physical
position of the community itself. A virtual community can exist solely in the
Internet or it can also be rooted to some place or a specific location. Nguyen et al.
(2008) also make a distinction between a place-based and a location-based virtual
community. According to Nguyen et al. (2008), a place is used for referring to a
meaningful location for humans, whereas a location is merely a point in the
coordinate system. In order to improve the clarity of this thesis, the term location-
based is adopted because it is still more established in the literature.
When considering the manner how communities are established, they are
typically either member-initiated or organisation-sponsored (Porter 2004).
Member-initiated communities can be further categorised into social and
professional communities based on their relationship orientation. On the same
basis, organisation-sponsored communities can be further categorised into
commercial, non-profit and government communities.
Commercial virtual communities, in particular, have been paid a great deal of
attention by the researchers of mobile and electronic commerce during the last
decade. For instance, Lechner & Hummel (2002) studied the business models for
virtual communities that are established using P2P technologies, Schubert &
Hampe (2005) examined business models for mobile virtual communities in
general, whereas Gupta & Kim (2004) investigated the potential applications of
virtual communities in both electronic and mobile commerce. According to Gupta
& Kim (2004), the potential benefits of virtual communities are versatile. For
instance, virtual communities provide a cost-effective way to communicate and
share knowledge, improve effectiveness of marketing, and enhance trust among
the customers.
27
Based on work conducted by Porter (2004), Tasch & Brakel (2004), Morr &
Kawash (2007) and Nguyen et al. (2008), Fig. 2 presents the topology of virtual
communities.
Fig. 2. The topology of virtual communities.
For the purpose of clarity, the term group is hereafter used in this section for
referring to any group, whether the group is a community or not. The reason for
this is that the systems presented in Section 2.1.2 could be all used for managing
groups, but not all of them for managing user communities. In some of these
systems, groups may, for instance, be established without users being aware of
this, which render these particular systems unsuitable for managing user
communities. To also clarify the terminology used in Papers I-III, it is important
to emphasise that in Paper I, all kinds of P2P GMSs are examined, but in Papers
Virtual community
Partially virtual Virtual
Degree of virtualisation
Stationary Mobile
Communication technology
Location-independent Location-based
Position
Member-initiated Organisation-sponsored
Establishment
Social Commerical
Relationship orientation
Professional Non-profit
Government
28
II and III, the focus is solely on using a specific type of a P2P system for
managing user communities.
2.1.2 Characteristics of a group management system
Today, GMSs have many applications. The purpose of a GMS can be, for instance,
social or professional networking, real-time group communication, sharing and
rating of multimedia content, as well as provisioning of collaborative services
(Boyd & Ellison 2007). In the given examples, groups are usually member-
initiated or organisation-sponsored. As a consequence, the user is aware of the
fact that the groups exist as an essential part of the service. Especially in P2P
networks, a GMS may also serve more technical purposes that are not directly
visible to the user. In these cases, the groups are often dynamically established by
the GMS in order to improve, for instance, the search performance of the service
(Agostini & Moro 2004, Li & Vuong 2006). Agostini & Moro (2004) emphasise
that the existence of a group and its members is often unknown to the user.
Because of these two different types of GMSs, it is important to examine the
common characteristics of a GMS from both technical and service viewpoints
(Tasch & Brakel 2004).
In the context of P2P networking, Gong (2001) introduces a group as a virtual
entity that implements a set of group protocols. Hence, from the technical
perspective, a GMS is comprised of multiple group management procedures. It is
important to note that the basic functionality provided by the group management
procedures is similar regardless of the chosen implementation technology or the
fundamental purpose of a GMS. A concise definition of group management could
not be found in the literature, but by completing the perceptions of Gong (2001)
and Bottazzi et al. (2008), the following technical definition was formulated:
group management refers to a set of procedures that enable establishing,
configuring, indexing, discovering, joining, leaving and removing groups. As a
result, a GMS is a system that implements, at least, part of these group
management procedures.
From the service perspective, a GMS provides a set of services that enable
interaction between the members of a group, and possibly between different
groups. Tasch & Brakel (2004) describe this set of services as multiple different
communication spaces that are populated by users. In addition to Facebook, well-
known examples of GMSs include LinkedIn that provides a professional network
for technology experts, MySpace that provides a meeting place for fans of social
29
entertainment, and Twitter that provides a real-time information sharing network
(Boyd & Ellison 2007). It is important to note that the services provided by these
GMSs rely on the defined set of group management procedures. For instance, a
chatting session with a particular friend requires discovering the group of friends
(i.e. friend list) that are online at that moment.
2.1.3 Technologies for group management systems
It should be emphasised that not only the purpose, but also the deployment
environment dictates what kinds of technologies are feasible for implementing a
GMS. The communication model and the group management procedures of a
GMS can be implemented using CS, MANET or P2P technologies, as well as
some combination of the previous. The choice between the alternatives depends
on varying criteria such as the predicted number of users, the mobility of users,
capabilities of users’ devices, as well as the available communication
infrastructure to mention a few examples. All three communication technologies
are presented in this section, but the focus is especially on P2P technologies in
order to provide the theoretical foundation for Papers I-III in this thesis.
Client-server networks
The communication model based on the CS architecture consists of a server and
one or multiple clients. A server acts as the provider of a service, whereas clients
act as users of that service. The clients are able to discover any service using a
Uniform Resource Locator (URL). A URL is comprised of the used
communication protocol, the name of the domain and the path to a particular
resource. With the CS architecture, the clients are lightweight compared to the
server. In the context of GMSs, all group management procedures are
implemented on the server using a variety of database structures that are, however,
outside the scope of this thesis.
All recently introduced GMSs (mainly SNSs) that rely on the CS architecture
are based on Web technologies (Boyd & Ellison 2007). With Web technologies,
the server hosts an additional component called a Web server that can be accessed
by the clients using a Web browser. Thanks to the rapid emergence of
smartphones and the fast wireless network interfaces, not only stationary
computers with a fixed Internet connection, but also high-end mobile devices are
capable of hosting a Web server (Srirana et al. 2006). The communication
30
between the Web server and the Web browser takes place using Hypertext
Transfer Protocol (HTTP) that operates over reliable Transmission Control
Protocol (TCP) (Kozierok 2005). The contents of a Web-based service are
exchanged as Hypertext Markup Language (HTML) documents that are
interpreted and visualised by the Web browser. The access to the database
structures on the server can be implemented with Web technologies, for instance,
using a scripting language called PHP. With PHP, HTML documents can be
dynamically created on run time.
The Web-based user interfaces of GMSs are today very often based on
Asynchronous JavaScript and XML (AJAX). AJAX is not a new technology, but
a combination of existing technologies used together in a novel way (Paulson
2005). With AJAX, a Javascript-based engine is run on the Web browser that is
capable of requesting content from the Web server silently in the background. As
a result, the interactivity and responsiveness of a Web-based service is
considerably improved.
The main advantages of using CS networks in GMSs are typically low
latency access to data, strong data security, good robustness against churn, and
low stress on clients which is especially important for mobile devices (Leibnitz et
al. 2007, Balasubramanian et al. 2009). The usual disadvantages include the
single point of failure, low scalability, a need for expensive hardware
infrastructure as well as high maintenance and administrative costs (Leibnitz et al.
2007). Some of these shortcomings, however, can be removed by using server
clustering techniques (Schroeder 2000). For instance, in Facebook and Twitter,
the issues related to the single of point of failure and low scalability have been
eliminated by using a distributed database called Apache Cassandra that is
capable of efficiently replicating all data within the established server cluster. As
a result, the speed of data access increases when new hardware is added.
Mobile ad hoc networks
A MANET is comprised of mobile nodes that are capable of dynamically self-
organising into arbitrary and temporary network topologies without any pre-
existing communication infrastructure (Chlamtac et al. 2003). Mobile nodes are
connected to each other using various wireless network interfaces such as
Bluetooth (BT) or WLAN. Because of the different wireless network interfaces
and their transmission capabilities, multi-hop protocols are usually needed for
enabling connectivity between mobile nodes that are not necessarily within each
31
other’s wireless range (Chlamtac et al. 2003, Adibi & Erfani 2006). Since no
default routers exist, each mobile node assumes the role of a router and forwards
other nodes’ packets in the established network topology. From the viewpoint of
security, any mobile node on the routing path can be potentially hostile. There
exists a large variety of routing protocols for MANETs (Yu et al. 2005, Adibi &
Erfani 2006), which are, however, not within the scope of this thesis.
In MANETs, all group management procedures need to be implemented in a
distributed manner. This is a challenging task, because the network topology is
changing dynamically, and thus, the correct routing information is difficult to
maintain (Adibi & Erfani 2006). This complicates, for instance, implementing
consistent indexing of group members. Without any centralised elements, the
indexing of group members would need to be implemented locally by each
mobile node according to their current knowledge of the network topology. Using
clustering-based routing schemes in MANETs, some well-connected nodes could
assume the role of a cluster leader (Yu et al. 2005). Because of their central
position in the network topology, the leader of each cluster could take the
responsibility for implementing the group management procedures. An example
of a MANET-based GMS that use clustering is presented by Bottazzi et al. (2008)
who cluster mobile nodes according to the similarity of their interests and goals.
The main advantage of using MANETs in GMSs is that no pre-existing
communication infrastructure is needed. In addition, the ability to exploit the
variety of wireless network interfaces enables the majority of mobile devices to
participate in a MANET. The main disadvantages are low security and the
difficulty of maintaining the network topology, which usually leads to rather low
performance, robustness and scalability. From the viewpoint of GMSs, this
complicates the implementation of group management procedures. However, it
should be noted that some of the group management procedures could also be
implemented using CS networks if several mobile nodes can act as gateways to
the Internet (Chlamtac et al. 2003).
Peer-to-peer overlay networks
An overlay network can be defined as a network that is established on top of
another network. In the case of P2P technologies, a P2P overlay network is
formed by a collection of logical links that connect nodes to each other at the
application layer (Androutsellis-Theotokis & Spinellis 2004). Each of these
logical links then corresponds to one or more physical links between the nodes.
32
Similar to MANETs, each message is propagated through multiple nodes that may
potentially be hostile. This, of course, has a negative effect on the security of P2P
systems (Urdaneta et al. 2011). From the viewpoint of communication
infrastructure, P2P overlay networks fall somewhere between CS and MANET
networks. P2P overlay networks rely on the existing IP-based networks, but do
not require any dedicated hardware components.
In the literature, P2P systems have been categorised in multiple ways, for
instance, according to the chronological appearance of different P2P system
architectures (Harjula et al. 2004) or the specific applications of a P2P system
(Brands & Karagiannis 2009). Androutsellis-Theotokis & Spinellis (2004) explain
that the diversity of the definitions results from the chosen level of broadness. In
their fundamental study of P2P systems, Androutsellis-Theotokis & Spinellis
(2004) reconcile the different viewpoints to P2P systems and define them as
“distributed systems consisting of interconnected nodes able to self-organise into
network topologies with the purpose of sharing resources such as content, CPU
cycles, storage and bandwidth, capable of adapting to failures and
accommodating transient populations of nodes while maintaining acceptable
connectivity and performance, without requiring the intermediation or support of
a global centralized server or authority”. Although the definition is self-
explanatory, understanding the P2P networking requires further examination of
the architecturally distinct approaches to implementing P2P systems.
From the architectural perspective, P2P systems can be categorised according
to the type of the index and the type of the network structure (Lua et al. 2005,
Brands & Karagiannis 2009). A P2P system can use either centralised, distributed
or hybrid indexing for locating nodes and resources in the P2P network. In
centralised indexing, the index is stored by one or more centralised servers. A
common example a P2P system that uses centralised indexing is the original
version of BitTorrent. In distributed indexing, the index is distributed to all nodes
in the P2P network and stored in a collaborative manner. An example of a P2P
system that relies on distributed indexing is Gnutella v0.4. In hybrid indexing, the
index is stored by a subset of nodes that are usually called super-nodes. Super-
nodes typically possess high capabilities and are responsible for maintaining the
index on behalf of ordinary nodes connected to them (Yang & Garcia-Molina
2003). An example of a P2P system that uses hybrid indexing is Skype (Baset &
Schulzrinne 2004).
Based on the type of the network structure, P2P systems can be further
categorised as unstructured, structured or a combination of unstructured and
33
structured. In unstructured P2P systems, nodes are organised in a random overlay
topology that emerges in the course of time. In unstructured P2P systems, nodes
and their shared resources are located using search methods that are either based
on flooding, random walks or expanding-ring search (Lua et al. 2005). A search
query is typically propagated until a specific Time-To-Live (TTL) parameter has
decreased to zero. As a consequence, a search query usually reaches only a part of
a P2P network, and thus, deterministic resolution of queries cannot be guaranteed.
An example of an unstructured P2P system is Gnutella v0.4.
In structured P2P systems, the structure of the overlay topology is tightly
controlled. In other words, each node and resource is mapped to an identifier that
has an unambiguous location in the overlay topology (Androutsellis-Theotokis &
Spinellis 2004). Apart from a few exceptions such as Mercury (Bharambe et al.
2004), all structured P2P systems are based on DHT algorithms. In DHT-based
P2P systems, nodes and resources are mapped into a logical numeric address
space. Within the address space, each node and resource is located using a key
that is calculated with a DHT-specific mathematical (usually hash) function (Xu
et al. 2003). If a key to a node or a resource is known, deterministic resolution of
search queries can be guaranteed. This is crucial when P2P overlay networks are,
for instance, used for managing user communities, managing time critical
information, provisioning of mobile services and/or storing specific items of the
service data. Without deterministic resolutions of queries, these different types of
resources could not be reliably found. This would result, for instance, in
expiration of time critical information such as location or in malfunctions of a
service. Popular DHT algorithms are Kademlia, Chord, CAN (Ratnasamy et al.
2001), Pastry (Rowstrom & Druschel 2001) and Tapestry (Zhao et al. 2004).
P2P systems that consist of both unstructured and structured overlay
topologies are typically based on hybrid indexing. In Skype, for instance, the
communication between the super-nodes is structured, whereas the
communication between a super-node and its ordinary nodes is unstructured.
In addition to the previously presented P2P systems, there are also more
complex P2P systems that are composed of multiple connected overlay networks
(e.g. Xu et al. 2003 Artigas et al. 2005, Li & Vuong 2006). In these P2P systems,
it is typical that there exists one overlay network that is populated by all the nodes
in the P2P system, whereas the rest of the overlay networks are comprised of
some subset of nodes. As a result, the type of the index and the type of the
network structure are tightly coupled and cannot be examined separately.
Commonly, the overlay topologies are based on DHT algorithms, and thus, these
34
P2P systems are usually called hierarchical DHTs. For instance, Xu et al. (2003)
introduce HIERAS, a layered P2P system, where nodes are arranged into different
levels of overlays based on their network latencies. This is done in order to
optimise the routing performance. Another P2P system called Cyclone is
presented by Artigas et al. (2005) who connect multiple overlays in a tree-like
conceptual hierarchy. Although several P2P systems that are composed of
multiple connected overlay networks exist, they have not been experimented for
managing location and/or activity-based user communities and have not been
evaluated using real-life prototype implementations. At least, in the case of DHT
algorithms, Kato & Kamiya (2007) found out that the results gained using
simulations may differ significantly from the measurements conducted with real-
life prototype implementations. Furthermore, Hautakorpi & Camarillo (2007)
pointed out that the suitability of a DHT algorithm depends on the deployment
environment and the specific application case. The PCMS, evaluated in Papers II
and III, is a P2P system that is based on hierarchical DHTs.
In addition to the type of the index and the network structure, Brands and
Karagiannis (2009) further categorise P2P systems based on the type of the
application, deployment, standardisation and security mechanisms. However, this
extensive set of categories is too vast to be examined in this thesis.
In P2P systems, the implementation of group management procedures is also
dependent on the selected type of the index and the structure of the network. With
centralised indexing, group management procedures are straightforward to
implement by the centralised servers (Krishnamurthy et al. 2001). In distributed
indexing, group management procedures are typically implemented either using
tree-like structures on top of the overlay topology (Castano & Montanelli 2005,
Castro et al. 2002) or locally by each individual node (Das et al. 2009). In hybrid
indexing, group management procedures are typically implemented by super-
nodes as, for instance, in JXTA (Gong 2001). In P2P systems that are composed
of multiple connected overlay networks, group management procedures are an
integrated part of the P2P system itself.
Advantages of using P2P technologies in GMSs are slightly challenging to
generalise because of the multiple architecturally distinct implementation
approaches. However, P2P networks have good scalability, mediocre robustness
against churn, and very low maintenance and administrative costs. The common
disadvantages are low security and mediocre performance, at least, when
compared to the CS networks. From the standpoint of mobile nodes, energy
consumption may also become an issue. This is due to two facts: first, the
35
maintenance of the network topology, especially in DHT-based P2P systems,
requires a fairly large amount of messaging; second, mobile nodes using a cellular
data connection such as 3G are usually behind network address translation (NAT)
or firewall (FW) devices that need to be traversed (Haverinen et al. 2007). NATs
and FWs have timers that are used for removing idle connections, and thus,
mobile nodes are required to send frequent keep-alive messages in order to keep
them reachable from the Internet. One solution to this problem is to give the
mobile nodes a public Internet protocol (IP) address. This, however, is not
possible as long as IPv4 is the dominating protocol in the Internet, because the
pool of unallocated IPv4 addresses was depleted already during the year 2011.
The IPv6, however, has a significantly larger address pool, and with public IPv6
addresses, the issue of the NAT traversal, at least, would be solved.
The presented communication models are not mutually exclusive and can be
used in a collaborative manner. For instance, the CS technologies could be used
for tasks that require a high level of security such as authentication, the P2P
technologies could be used for communication and resource sharing, whereas the
MANET technologies, in turn, could be used whenever communication
infrastructure is not available or when the costs for the Internet access are
considered too high (Chlamtac et al. 2003). In Table 1, a rough conceptual
comparison of CS, P2P and MANET-based communication models is presented
in terms of performance, robustness, scalability, security, suitability for mobile
nodes, independence of communication infrastructure and independence of
maintenance and administration.
Table 1. A conceptual comparison of CS, P2P and MANET communications models.
Comparison criteria Communication model
CS P2P MANET
Performance high medium low
Robustness high medium low
Scalability medium high low
Security high low low
Suitability for mobile devices high medium high
Independence of communication infrastructure low medium high
Independence of maintenance and administration low high high
36
2.2 Convergence of technologies in mobile environments
The heterogeneity of wireless network interfaces, mobile OSs and communication
networks introduces a respectable challenge for today’s mobile software
developers (Yoneki & Bacon 2007). However, the integrated use of different
technologies can be facilitated using a middleware solution. In this section, the
benefits of mobile middleware are investigated together with the versatile role of
context information. Furthermore, the integrated use of P2P and Web technologies
in mobile environments is examined. The goal of this section is to provide the
theoretical background for Papers IV and V in this thesis.
2.2.1 Benefits of mobile middleware
Middleware is an umbrella term that comprises of “a set of services, APIs, and
management systems, which support distributed, networked computing
environment” (Yoneki & Bacon 2007). Consequently, a mobile middleware is a
special case of middleware that is run on mobile devices. What makes the
development of mobile middleware especially challenging is the scarcity of
available resources in mobile environments (Tarkoma 2009). As a consequence, a
mobile middleware must be able to withstand delays that are usually caused by
the unreliable wireless network connections and the relatively low processing
capabilities of mobile devices. Added to this, the mobile middleware itself should
be lightweight in order to preserve the precious battery life of a mobile device.
Typically, a middleware is implemented as a software layer that resides above
the OS and the TCP/IP stack, but below the applications (Tarkoma 2009).
According to Yoneki & Bacon (2007), middleware can be categorised roughly
into two different types: application-specific upper middleware and resource-
specific lower middleware. Application-specific upper middleware provides a set
of services that are used for simplifying the application logic and for enhancing
the interoperability between the applications. Resources-specific lower
middleware, in turn, is used for managing particular resources in a distributed
network environment. A mobile middleware (i.e. P2P Daemon), introduced in
Paper V, implements some of the features from both the application-specific
upper middleware and the resource-specific lower middleware. P2P Daemon
provides a service that enables Web-based services running in a Web browser to
utilise the available P2P networks, and at the same time, P2P Daemon is required
to participate in these P2P networks as one of the nodes.
37
In Fig. 3, a generic mobile middleware stack is illustrated (Yoneki & Bacon
2007). Because an exact composition is dependent on the purpose of a mobile
middleware and its deployment environment, Yoneki & Bacon (2007) provide a
generic set of services that represent the typical building blocks used in mobile
middleware solutions.
Fig. 3. A generic mobile middleware stack (modified from Yoneki & Bacon 2007).
The benefits provided by a mobile middleware are diverse when examined from
the viewpoint of different actors in the value chain of mobile business (Yuan &
Zhang 2003, Tarkoma 2009). For end-users, a mobile middleware as such is not
visible, but it provides indirect benefits through the wider range of supported
applications. For device manufactures, a mobile middleware enables introduction
of extended features that interface with the device drivers. For Internet service
providers, a mobile middleware enables monitoring and administration of the
network environment. For platform providers, a mobile middleware provides
interoperability with different mobile OSs. Finally, for application service
providers, a mobile middleware enables easier and faster application development.
Overall, the use of a mobile middleware reduces costs and also shortens the time
required for introducing new devices and services to the market.
Open APIs
Network architecture
Middleware generic service elements
Operating system
Mobile middleware communication
Client-Server P2P/Grid MANET Wireless sensor network
Resource managementData management
Event notificationService discovery
Directory
Quality of ServiceReal-time
Security
Context management
38
The number of existing mobile middleware solutions is vast. In the following,
illustrative examples of MANET and P2P-based mobile middleware solutions are
introduced. The examples include Proem (Kortuem 2002), STEAM (Meier &
Cahill 2010), PnPAP (Harjula et al. 2004) and MyNet (Kalofonos et al. 2008).
Kortuem (2002) introduces Proem that is a mobile middleware for developing
and deploying P2P applications in MANETs. The architecture of Proem consists
of three parts that are an application run-time environment, a set of middleware
services and a protocol stack. The middleware services include the presence
manager, the profile manager, the data space manager, the community manager,
the peer database and the event bus that enables event-based communication
among the P2P applications. Another example of a mobile middleware for
MANETs is presented by Meier & Cahill (2010) who introduce STEAM, an
event-based middleware for collaborative applications. The architecture of
STEAM comprises of four parts that are the event service nucleus, the proximity
discovery service, the location service and the proximity-based group
communication service. As part of STEAM, Meier & Cahill (2010) introduce
techniques for location-independent announcement and subscription coupled with
location-dependent filtering and event delivery.
Harjula et al. (2004) describe a mobile middleware called PnPAP that enables
dynamic selection of a P2P protocol, a session management protocol and a
wireless network interface. Configuration and run-time optimisation of PnPAP
can be done using a lightweight interpretable state machine. Another example of a
mobile middleware relying on P2P technologies is MyNet that provides secure
personal and social networking services (Kalofonos et al. 2008). MyNet enables
user-generated services and content to be accessed and shared in real-time,
directly from users’ mobile devices. The shared services can also be implemented
using Web technologies in which case they can be run on any Web browser.
MyNet is built on top of a distributed communication platform called Unmanaged
Internet Architecture (UIA) (Ford et al. 2006).
2.2.2 Versatile role of context information
“Context is all”
Margaret Atwood 1985
Context as a term can be very vague and ambiguous if not defined properly. This
is a result from the fact that practically everything around us is happening in a
39
certain context (Chen & Kotz 2000). Added to this, context has diverse meanings
in different areas of computer sciences. In mobile computing, the most
established definition is given by Dey et al. (2001) who define context
information as “any information that can be used to characterize the situation of
entities (i.e. whether a person, place, or object) that are considered relevant to the
interaction between a user and an application, including the user and the
application themselves. Context is typically the location, identity, and state of
people, groups, and computational and physical objects”. The definition is precise,
and at the same time, generic enough in order to cover all types of context
information and application scenarios. Furthermore, the definition takes into
account that an entity can also be a group of individuals (Ferscha et al. 2004).
In the literature, there exists a large variety of attempts to categorise context
information (e.g. Chen & Korz 2000, Dey et al. 2001, Krogstie et al. 2003).
Introducing multiple different categorisations of context information, however, is
not fundamental to this thesis, and thus, the widely accepted categorisation given
by Dey et al. (2001) is adopted and presented in more detail.
Dey et al. (2001) classify context information into four categories that are
identity, location, status and time. Identify refers to a unique identifier given to an
entity. In the used namespace, each identifier must be unique so that the
applications are able to unambiguously identify the corresponding entities. For
instance, in DHT-based P2P systems, each node is given a peer-ID that uniquely
identifies both the node and its location in a global namespace (Baset et al. 2007).
In many cases, location is usually associated only with an exact position in a
two-dimensional space. However, location can also be expanded to include
meaningful places, orientation, elevation and spatial relationships between entities.
In order to provide a standard software engineering model for location-based
ubiquitous systems, Hightower et al. (2002) introduce a seven-layered location
stack. The layers are in a bottom-up order sensors, measurements, fusion,
arrangements, contextual fusion, activities and intention. In the location stack, the
raw data is first gathered from sensors and transcribed into canonical
measurement types. This measurement data is then combined to identify the
representation and the orientation of entities that are arranged into a coordinate
system. Finally, location data is merged with other non-location contextual
information in order to identify the activities and the intensions of a user.
The characteristics of an entity are identified with status. Depending on the
nature of an entity, status can include multiple different kinds of context
information. First, if an entity is a person or a group of people, status can refer to
40
physiological factors such as mood or to an activity such as chatting with a friend.
Second, if a place, status can be the current air temperature or the noise level.
Third, if a computing system, status can include the current load on the central
processing unit (CPU) or the available wireless network bandwidth. Finally, time
is often used in association with the other pieces of context information. When
expressed as a timestamp or a timespan, time indicates when some other piece of
context information has been discovered or is to be considered relevant.
Context information can be gathered from multiple different sources. For
instance, new smartphones are full of sensors that are able to determine the
location of the user, the current air temperature, as well as the orientation and
acceleration of the device itself to mention only a few examples. Some pieces of
context information such as status can also be added manually by the users, but in
an effective context-aware system, the gathering of context information should, of
course, be as automatic as possible. In order to efficiently manage context
information in a distributed network environment, common methods for
modelling and presenting as well as for distributing context information are
required.
Strang & Lindhoff-Popien (2004) introduce six different approaches to
modelling context information. The approaches are key-value, markup scheme,
graphical, object-oriented, logic-based and ontology-based models. Key-value
models are very easy to manage, but they cannot be structured in a sophisticated
way to enable efficient context retrieval algorithms. In markup scheme models,
the context information (typically profiles) is expressed as a hierarchical data
structure consisting of markup tags with attributes and content. The
implementation of markup scheme models is usually based on Extensible Markup
Language (XML). For instance, the personal music profiles, introduced in Paper
VII, are expressed using a simple markup scheme model. In graphical models,
context information is presented using some graphical modelling language such
as Unified Modelling Language (UML) or Object-Role Modelling (ORM)
(Henricksen & Indulska 2006). In object-oriented models, the advantages of the
object-oriented approach, namely encapsulation and reusability, are exploited for
managing the dynamicity of context information in a ubiquitous network
environment. The details of processing the context information is encapsulated
within the objects, and thus, hidden from the other components. In logic-based
models, context information is defined as facts, expressions and rules. By
reasoning, expressions and facts can be derived from conditions according to a set
of rules. Finally, ontology-based models describe context information as concepts
41
and relationship between them. This makes ontology-based models especially
suitable for systems that manage context information used in our daily life.
Distribution of context information is typically implemented according to two
different approaches. First, in a push-based approach, context information is
pushed to an entity that is responsible for storing and distributing context
information. This entity can either be a centralised server or a distributed data
structure, as in Skype (Based & Schulzrinne 2004). Second, in a pull-based
approach, context information is stored locally by each device and is pulled when
needed. The context information can either be retrieved directly from a device or
through a gateway entity, as illustrated by Järvinen et al. (2009a). Context
information can be made available, for instance, as a Web-based service. In Paper
V, the push-based approach was adopted, because it enables easier retrieval and
delivery of community context information.
For carrying both presence and location information, Internet Engineering
Task Force (IETF) has introduced an XML-based data format called Presence
Information Data Format Location Object (PIDF-LO) (Sugano et al. 2004,
Winterbottom et al. 2009). It is important to notice that the IETF term presence
can be seen equivalent with the term status used by Dey et al. (2001) in their
categorisation of context information. For each piece of presence and location
information, PIDF-LO also encompasses data fields for the entity and the time
information. It should be noted that PIDF-LO only defines the data format for
carrying context information, but does not concern with the data format of the
context information itself. In Paper V, multiple PIDF-LO documents were
bundled together for enabling efficient delivery of community context
information.
There are many application scenarios for context information. However,
without contradictions, the most common category of context-aware applications
is related to navigation (Dey et al. 2011). For instance, Ohtonen et al. (2006)
introduce NaviP2P, a group-based mobile P2P navigation application. In NaviP2P,
locations of the user and her group member are shown on a map interface. By
selecting a group member on the map interface, an application session such as
chat can be instantly proposed. NaviP2P is built on top of the previously
presented PnPAP middleware. Another common category of context-aware
applications is related to proximity sensing (Chen & Kotz 2000). For instance,
Väänänen-Vainio-Mattila et al. (2010) present TWIN, a MANET-based social
networking service, which offers applications for social presence, multimedia
42
sharing and group communication. By using WLAN in an ad hoc mode, a user is
able to sense her surroundings in order to discover other users in the proximity.
There also exists a variety of context-aware applications that automatically
perform particular operations when certain conditions have been met. This
category of context-aware applications is related to event triggering (Gu et al.
2004). Illustrative examples include a location-based reminder service and a fall
detection service that can automatically contact a medical unit (Noury et al. 2007).
Finally, one more important category of context-aware applications is related to
automatic configuration (Chen & Kotz 2000). For instance, Chen & Kotz (2000)
describe a context-aware mobile service that can adjust the font size and the
brightness of the screen according to the status of a user.
2.2.3 Integrated use of P2P and Web technologies
Today, the research on the integrated use of P2P and Web technologies in fixed-
line Internet is dominated by three different approaches. First, there exists a huge
number of proposals for replacing the centralised Web Services repository such as
Universal Description, Discovery and Integration (UDDI) with P2P technologies
in order to enhance the scalability and robustness of Web Service provisioning
(e.g. Li et al. 2005, Yu et al. 2007, Birman et al. 2009). Second, P2P technologies
have been used for caching and replication of Web content (e.g. Wanlei & Casey
2008, Iwamaru et al. 2009). The benefits of this approach are reduced access
latency to the Web content and lighter load on Web servers. Third, in some
proposals, access to P2P overlay networks has been enabled for Web-based
services running on a Web browser in order to realise a scalable and distributed
Web server architecture (e.g. Deshpande et al. 2007, Terrace et al. 2008). In this
approach, the communication between a Web-based service running on a Web
browser and P2P overlay networks is usually implemented in the source code of
the Web-based service itself and as a plug-in to a Web browser. Alternatively, the
source code of a Web browser could also be modified, but this would severely
decrease the interoperability of the implementation. In Paper V, the
communication between Web-based services and P2P overlay networks was
implemented using a specific software library in the Web-based services and a
plug-in in the Web browser.
As a research area, the integrated use of P2P and Web technologies in mobile
environments is rather new. This is a result mainly from the fact that until recently,
the capabilities of mobile devices have not been sufficient for participating in P2P
43
overlay networks. As in the fixed-line Internet, P2P technologies can also be used
for the provisioning of mobile Web-based services. For instance, Gehlen & Pham
(2005) propose a Service-Oriented Architecture (SOA) for providing mobile Web
Services in a distributed network environment. A SOA-based realisation is
proposed for both P2P and MANET-based networks. In their architecture, Web
Services are published as XML documents using Web Service Description
Language (WSDL). Srirama & Prinz (2007) introduce an integration framework
also for provisioning of mobile Web Services. The framework is based on
Enterprise Service Bus (ESB) technology that combines Web Services, messaging
middleware, intelligent routing, and transformation. The ESB acts as an
intermediary between the Web Service clients in the Internet and the mobile
nodes in a P2P overlay network that is based on JXTA. Finally, Hassan (2009)
proposes another mobile Web Service architecture based on an intermediary
gateway and a JXTA-based P2P overlay network. The intermediary gateway is
used for connecting the Web Service clients and providers in the JXTA-based P2P
overlay network.
P2P technologies have also been used for implementing a GMS for Web-
based social networking services. For instance, Kalofonos et al. (2008) present
MyNet, a P2P-based social networking platform that supports provisioning of
Web-based services for user groups. An example application implemented for
MyNet is a Web-based content browsing application that enables access to the
contents in the user’s personal devices. Another example, in which P2P
technologies have been used for implementing a GMS for a Web-based service, is
collaborative browsing of rich Web pages that are too complex and large to be
displayed efficiently on a single mobile device (Xiao et al. 2008). In their work,
Xiao et al. (2008) introduce an adaption engine that splits a rich Web page into
multiple small blocks. The adaption engine is implemented in a distributed
manner using a layered P2P structure. Each layer is established as an unstructured
P2P overlay network that relies on centralised indexing. In each layer, the most
capable mobile node acts as a master (i.e. a super-node) and takes the
responsibility for allocating the blocks to the other nodes connected to it. As
supported by the given examples, the research on the integrated use of P2P and
Web technologies in mobile environments has quickly gained momentum.
However, the existing research still lacks a general-purpose service framework,
where all functionalities of a P2P system could be seamlessly used by a Web-
based service running in a Web browser of a mobile device.
44
In order to summarise the benefits of the integrated use of P2P and Web
technologies in mobile environments, the following characteristics can be brought
forth. Use of Web technologies enables easy and fast development of mobile
services that are independent of the mobile OS. P2P technologies, in turn, enable
easy and scalable implementation of user communities. Furthermore, the use of
P2P technologies increases the scalability and robustness of the provisioning of
Web-based services.
2.3 Evaluating user acceptance of information systems
Understanding the users’ perceptions of a service, and its implications on the
usage intentions, is a challenging task for the researchers of information systems
(ISs). This results from the fact that the users’ perceptions of a service are not
merely affected by technical standards, but also by complex human factors that
can range from attitudinal to aesthetical elements. In this section, a user-centred
approach to the development of ISs is presented together with theoretical models
that can be used for assessing the acceptance of novel information technologies.
2.3.1 Defining user experience
Before going into the details of user-centred design (UCD) and the models of user
acceptance, it is important to distinguish two prominent concepts, namely quality
of service (QoS) and quality of user experience (QoE), from each other.
During the last two decades, QoS has become a very commonly used, but
also rather ambiguous, acronym in the research of communication networks.
International Telecommunication Union (ITU) (2009) defines QoS as “the totality
of characteristics of a telecommunications service that bear on its ability to satisfy
stated and implied needs of the user of the service”. Despite the efforts to
establish QoS as a measure of user experience (UX), the term has been primarily
used in a technical context (Reichl 2007, Kilkki 2008). Although the network-
related QoS parameters, such as bit rates, delays and packet loss rates, might be
crucial for the fluent operation of an application, UX in a service is also affected
by more human factors (Hassenzahl & Tractinsky 2006).
The concept of UX was adopted into the research of ISs to increase the
understanding on “how people interact with products, other people and the
resulting emotions and experience that unfold” (Forlizzi & Battarbee 2004). The
complex phenomenon of UX has also been recently compacted by International
45
Organization for Standardization (ISO) (2009) who defines UX as “a person’s
perceptions and responses that result from the use or anticipated use of a product,
system or service”. In order to have a common denominator for the evaluation of
UX, the concept of the quality of user experience (QoE) was introduced. The
relationship between the acronyms of QoS and QoE is emphasised by
Muhammad et al. (2006) who state that the aim of the network and services is to
guarantee maximum user satisfaction (i.e. QoE), while the network quality (i.e.
QoS) is needed for achieving that goal effectively. Whereas the evaluation of QoS
is a relatively straight-forward task using the previously mentioned network-
related parameters, the approaches to evaluating QoE are more diverse and
isolated due to the complex nature of UX itself (Forlizzi & Battarbee 2004). In
this thesis, QoE was not particularly evaluated, but rather its consequences on the
user acceptance of exemplary community-centric mobile services.
2.3.2 User-centred design
The idea of the UCD was first introduced by Norman & Draper (1986) who
emphasised the necessity of understanding the users’ needs in the development of
ISs. From the beginning, UCD has been strongly connected with the concept of
usability that is defined as “a measure how well the users are able to use the IS”
(Nielsen 1993). In the late 1980’s, the academic interest towards the UCD began
to increase, especially when it became evident that developing usability of ISs
was mainly seen as an activity carried out only in the end of the development
process (Johnson et al. 2007).
Based on the original mindset brought up by Norman & Draper (1986), the
purpose of UCD is, specifically, to guarantee that the users are able to use the ISs
to fulfil their tasks. This initial goal of UCD is solid and understandable, but lacks
practical details that are necessary for successful implementation of the mindset.
This shortcoming has raised a lot of criticism towards the UCD (e.g. Constantine
& Lockwood 2002), but at the same time, has also fuelled the development of
versatile UCD methodologies. As an outcome of two decades of academic
discussion and numerous practical use cases, UCD has evolved into “a process
that focuses on usability throughout the entire development process and further
throughout the IS life cycle” (Gulliksen et al. 2003). In this regard, UCD is now
seen as an integrated and continuous process composed of various user-centred
activities. These activities include, for instance, explicit user focus, active user
involvement, iterative and incremental systems development and participation of
46
a multidisciplinary development team (Gulliksen et al. 2003). In the development
of the two exemplary community-centric mobile services, introduced in Papers VI
and VII, some of the key activities of UCD were adopted.
Besides UCD, research efforts on the development of ISs have resulted in the
introduction of two other essential fields of research called usability engineering
and information technology acceptance. Usability engineering bears similarities
with UCD, because of its strong emphasis on retrieving continuous feedback from
the actual or potential users throughout the entire development cycle of ISs
(Nielsen 1993). Compared to UCD, usability engineering, however, has a
narrower scope, since its main focus is mainly on defining measurable usability
goals and evaluating these goals against the developed IS. Fulfilment of the
usability goals is commonly verified using a method called usability testing,
where representative users use an IS under observation of usability experts
(Kushniruk & Patel 2004). In usability testing, it is also common that the entire
interaction between the users and the IS is recorded and minutely analysed
afterwards. Although usability engineering has evolved as an independent field of
research, it easily fits as a part of a more comprehensive UCD process.
As a field of research, information technology acceptance approaches the
development of ISs from a different standpoint and considers usability only as
one of the factors that explain the reasons why some ISs are, or are not, accepted
by users (Venkatesh et al. 2003). Along with usability engineering, information
technology acceptance provides practical methods that can be used in different
stages of the UCD process and the IS life cycle. In the next subsection, some
widely applied models for information technology acceptance are summarised in
order to provide a theoretical foundation particularly for Paper VI in this thesis.
2.3.3 Models of user acceptance for information systems
“Essentially, all models are wrong, but some are useful”
George E.P. Box 1987
Research on information technology acceptance has its roots in three distinct
fields of research that are ISs, psychology, and sociology. During the last 50 years,
this multidisciplinary field of research has introduced several competing models
that aim to resolve the determinants that have the most significant effect on both
the use intentions and the actual use of an IS. It is common that each of the
models presents its own set of determinants that are, however, partially
47
overlapping (Venkatesh et al. 2003). Today, the most prominent models are: (1)
Innovation Diffusion Theory (IDT) (Rogers 1962, 2003), (2) Theory of Reasoned
Action (TRA) (Fishbein & Ajzen 1975), (3) Theory of Planned Behaviour (TPB)
(Ajzen 1991), (4) Technology Acceptance Model (TAM) (Davis et al. 1989), and
(5) Unified Theory of Acceptance and Use of Technology (UTAUT) (Venkatesh
et al. 2003). Three of the five models: IDM, TRA and TPB have not been
specifically developed for modelling user acceptance of information technology,
but have been successfully applied to this context as well.
Innovation Diffusion Model
In the early 1960’s, Rogers (1962, 2003) introduced a generalised innovation
diffusion model (i.e. IDT) through a synthesis of several theories on innovation
adoption behaviours. In general, the aim of IDT is to explain how innovations are
diffused in a social system, and how the members of this social system adopt
these innovations. Rogers describes diffusion as “a process by which innovation
is communicated through certain channels over time among the members of a
social system”. In IDT, the diffusion process is distinguished from the adoption
process. Whereas the diffusion process takes place within a social system, the
adoption process is more closely related to individuals.
Fig. 4. Innovation adoption curve.
According to Rogers, individuals can be classified into five different categories
based on their degree of innovativeness that has influence on their adoption
Time
Cum
ulat
ive
adop
tion
Innovators (2,5%)
Early adopters (13,5%)
Early majority (34%)
Late majority (34%)
Laggards (16%)
48
behaviour. Over the length of time, the cumulative adoption of an innovation
forms an S-shaped diffusion curve, as illustrated in Fig. 4.
In order to discover the reasons why some innovations are adopted and some
are rejected, Rogers (1962, 2003) identified five characteristics of an innovation
that have the most significant influence on adoption behaviours. These
characteristics include relative advantage, compatibility, complexity, trialability,
and observability. This original set of characteristics has been later modified and
extended, for instance, by Tornatzky & Klein (1982) and Moore & Benbasat
(1991) who also provided an instrument for examining, especially the adoption of
information technology innovations. Overall, IDT has been widely used in the
studies of information technology and systems adoption. For instance, Agarwal &
Pradat (1997) examined the adoption of WWW, Van Slyke et al. (2002) focused
on the adoption of group-based communication applications, and to conclude,
Prescott & Conger (1995) listed over 70 publications on information technology
and systems adoption that were published in 1984–1993.
Theory of Reasoned Action
In the mid 1970’s, Fishbein & Ajzen (1975) introduced TRA (Fig. 5), a very
general model used for predicting and explaining behaviour in a variety of
domains. TRA has its origins in social psychology which is interested in the
determinants of consciously intended behaviour (Davis et al. 1989).
Fig. 5. Theory of reasoned action (Fishbein & Ajzen 1975).
According to TRA, the actual behaviour of an individual is determined by her
intention to conduct the behaviour, which is further influenced by both the
individual’s attitude and subjective norm (Teo & Pok 2003). Attitude represents
an individual’s beliefs on the consequences of a particular behaviour and the
evaluation of the pleasantness of these consequences (Fishbein & Ajzen 1975). In
turn, the subjective norm is defined as an individual’s beliefs whether the people
Attitudetowards
behaviour
Subjectivenorm
Behaviouralintention Behaviour
49
that are important to her expect her to conduct or not to conduct that particular
behaviour, and whether she is motivated to comply (Teo & Pok 2003). Since TRA
is a general model, it does not specify what beliefs are relevant for a specific
behaviour. Consequently, researchers using TRA must first identify the beliefs
that are essential regarding the behaviour under examination (Davis et al. 1989).
Although used in variety of domains, TRA has received most attention from the
researchers of marketing that study customer behaviour (Sheppard et al. 1988).
Theory of Planned Behaviour
Despite its popularity, TRA was found to have difficulties when dealing with
behaviours in which people do not have volitional control (Teo & Pok 2003). To
overcome this specific shortcoming, Ajzen (1991) introduced TPB (Fig. 6), an
extended version of TRA, in which a third element called perceived behavioural
control was added as a determinant of both behavioural intention and actual
behaviour. Perceived behavioural control refers to an individual’s perception of
her resources, skills, and opportunities to conduct a particular behaviour, and
whether she considers the expected results of that behaviour important. According
to Teo & Pok (2003), perceived behavioural control has been discovered to have a
significant influence, especially on the use behaviour of information technology.
Fig. 6. Theory of planned behaviour (Ajzen 1991).
Besides the original version, TPB has also a decomposed variant (DTPB), in
which the three determinants: attitude, subjective norm, and perceived
behavioural control are further decomposed into an underlying belief structure
(Taylor & Todd 1995). Although DTPB is slightly more complex than TRA, TPB
or TAM (Taylor & Todd 1995), Teo & Pok (2003) emphasise that DTPB provides
Attitudetowards
behaviour
Subjectivenorm
Intention Behaviour
Perceivedbehavioural
control
Behaviouralbeliefs
Normativebeliefs
Controlbeliefs
50
a few significant advantages over its competitors. First, the underlying belief
structure of DTPB can be constructed of specific innovation characteristics that
are relevant to the particular behaviour under investigation. Second, DTPB
provides more explanatory power, and thus, a more thorough understanding of the
determinants that influence the use behaviour of information technology. Studies
of information technology and systems adoption that rely on TPB or DTPB are
vast in numbers. For instance, Ajjan & Hartshorne (2009) investigated the
acceptance of Web 2.0 technologies in academic environments, Lu et al. (2009)
examined acceptance of instant messaging in China, and Teo & Pok (2003)
explored the adoption of WAP-enabled mobile phones among Internet users. In
this thesis, DTPB was chosen as a research model for Paper VI.
Technology Acceptance Model
For modelling acceptance of ISs, in particular, Davis et al. (1989) introduced
TAM (Fig. 7). According to Taylor & Todd (1995), TAM can be considered as a
special case of TRA, in which behavioural intention, and further the actual use, is
influenced only by attitude that is determined by two beliefs: perceived usefulness
and perceived ease of use. Perceived usefulness is defined as an individual’s
perceptions whether using a particular IS will increase her performance within an
organisational context (Davis et al. 1989). In turn, perceived ease of use refers to
an individual’s perceptions whether using a particular IS will be free of physical
and mental efforts (Cheong & Park 2005).
Fig. 7. Technology Acceptance Model (Davis 1989).
TAM has been found to be successful in predicting the use of ISs (Taylor & Todd
1995), especially when applied in the later stages of the software development
process using a rather mature version of the system (Dillon 2001). It is important
Perceivedusefulness
Perceivedease of use
Behaviouralintention to use
Actualsystem use
51
to notice that using an immature version might produce skewed perceptions of
both usefulness and ease of use of the IS under examination.
TAM has been very popular among the researchers of ISs acceptance. To
demonstrate this, Legris et al. (2003) identified altogether more than 80 journal
papers published in 1980–2001, where TAM was used as a research model.
During the last 10 years, many TAM-based studies have also been conducted for
mobile ISs. For example, Shin (2007), Cheong & Park (2005), and Lee et al.
(2002) examined the acceptance of mobile Internet, whereas Pagani (2004)
focused on the adoption of the third generation mobile multimedia services.
Besides the original version of TAM, several modifications and extensions have
been proposed, for instance, by Shin (2007), Cheong & Park (2005), and
Venkatesh & Davis (2000). Thus far, the most prominent extension of TAM has
indisputably been TAM2 with more than two thousand citations found in Google
Scholar (Venkatesh & Davis 2000).
Unified Theory of Acceptance and Use of Technology
After the turn of the millennium, a new trend started to emerge from the ISs
research community. Instead of perfecting the existing models, more emphasis
was laid on an integrated approach, in which the features of two or more models
are combined (Yi 2006). One of the recent proposals has been provided by
Venkatesh et al. (2003) who introduced UTAUT (Fig. 8) that draws and combines
elements from eight different acceptance models building a unified view to
information technology acceptance. In UTAUT, four determinants: performance
expectancy, effort expectancy, social influence, and facilitating conditions have a
direct influence on behavioural intention and use behaviour. In addition, gender,
age, experience and voluntariness of use are considered as key moderators that
may further influence the four determinants and their interpretation.
During the recent years, UTAUT, too, has been applied by the research
community. For instance, UTAUT was used by Eckhardt et al. (2009) to study the
significance of social influence on the adoption of information technology, and by
Koivumäki et al. (2008) to discover how familiarity of mobile devices, the time
used for testing and individuals’ technology skills affect the acceptance of mobile
services. For instance, Koivumäki et al. (2008) found out that users’ perceptions
on a mobile service did not notably depend on the length of the test period, which
emphasises the importance of the first impression. It was also discovered that
skilled users tend to consider mobile services more useful and easier to use.
52
Fig. 8. Unified theory of acceptance and use of technology (Venkatesh et al. 2003).
Fuelled by the rapid development of mobile technologies, the user communities
that are established according to the shared location and/or activities of people are
becoming technically feasible. In order to succeed in the deployment of the novel
community-centric mobile services, more knowledge is required on the factors
that affect the user acceptance of this type of new mobile services. As supported
by the given examples in the section, user acceptance of Web-based mobile
services has already been paid a lot of attention in the research community of MIS.
However, according to the author’s best knowledge, studies on user acceptance of
community-centric services are very few (e.g. Van Slyke et al. 2002, Dholakia et
al. 2004) and mostly lacking the mobile aspect.
Performanceexpectancy
Effortexpectancy
Behaviouralintention
Usebehaviour
Socialinfluence
Facilitatingconditions
Gender Age Experience Voluntariness
53
3 Summary of the research contributions
In this chapter, the research contributions of the original publications are
presented in detail. Furthermore, it is discussed how each original publication
answers to the research questions proposed in this thesis. In Section 3.1, a
conceptual analysis of P2P GMSs is presented. In Section 3.2, the performance of
the PCMS is evaluated. In Section 3.3, the research issues related to the
implementation of the COMSE are addressed. In Section 3.4, two pilot services
built on the COMSE and user studies of these services are presented, and finally,
Section 3.5 discusses the significance of the results and the items for future work.
3.1 Conceptual analysis of P2P group management systems
The first research question in this thesis concerned efficient ways to establish
groups with different P2P system architectures. Paper I, entitled “P2P Group
Management Systems: A Conceptual Analysis” contains the contributions to this
research area and provides answers to the first research question. First, the
publication provides a concise definition of a P2P GMS and a description of P2P
group management procedures. Second, the publication illustrates how groups are
efficiently formed using different P2P system architectures, and analyses the
advantages and disadvantages of using each P2P system architecture for
implementing P2P group management. Both the conceptualisation and the survey
of P2P GMSs are unique attempts to provide a comprehensive understanding of
P2P group management.
Based on the conceptual analysis, P2P systems with hybrid indexing are the
most suitable for community members who are using battery-powered devices.
This is because the mobile nodes can assume a less demanding role of an ordinary
node. However, if a P2P system is almost entirely populated by mobile nodes,
some of them are required to assume a more demanding role of a super-node. In
this case, hybrid indexing loses its advantage. P2P systems that are composed of
multiple overlays have been rarely used for P2P community management, but as
they can provide a complete autonomy of groups, each group could be
implemented using any type of a P2P system that is chosen efficiently according
to the group-specific needs. Furthermore, the autonomy of groups inherently
improves security and fairness of group management procedures, because the
load originated from the activity within a particular group is not shared by other
groups.
54
The publication concludes by providing instructions for facilitating the task
of designing a P2P GMS. The instructions are summarised as follows:
1. Define the purpose why groups are needed.
2. Define the set of group management procedures that need to be implemented.
3. Estimate the characteristics of the intended usage environment.
4. Define the required qualities of the underlying P2P system architecture.
5. Consider how to implement the needed group management procedures.
6. Select an underlying P2P system architecture for implementing the defined
P2P GMS.
3.2 Performance evaluation of a P2P community management system based on two-level hierarchical DHT overlays
The second research question in this thesis deals with the implementation of a
robust and efficient community management system for mobile nodes using
structured P2P technologies. Papers II and III contain the contribution to this
research area and provide answers to the second research question. Paper II also
provides some evidence to the first research question. This is achieved by
providing an example of a GMS that is designed for user communities and
implemented using multiple connected P2P overlay networks.
Paper II, entitled “Improving Community Management Performance with
Two-level Hierarchical DHT Overlays” introduces the PCMS, a P2P community
management system that was originally developed using P2PP as a DHT-based
P2P protocol and Kademlia as a DHT algorithm. In the PCMS, a separate DHT
overlay is established for each community, resulting in multiple small overlays
that are indexed and discovered using a global DHT overlay (henceforth called
the main overlay). Additionally, the community member lists are indexed and
discovered in the main overlay. In comparison to a single large DHT overlay
where all the community-related activities take place, the PCMS provides the
following benefits:
– More efficient routing, because of the smaller community overlays
– Better security, because the activities of different community overlays are
separated from each other
– Fairer load distribution, because the community-related activities are carried
out only by the members of a specific community overlay.
55
If supported by the used DHT-based P2P protocol, nodes can assume either the
role of a peer or a client. A peer participates in the maintenance of the DHT
structure, whereas a client connects to a peer and merely uses the services
provided by a community through its peer. Each node takes part in the main
overlay, but it is beneficial that only some of these nodes assume the role of a peer.
This results from the fact that the messaging in the main overlay comprises only
of some publish and look-up operations which can be easily managed by a
relatively small number of peers. Instead, having a large overlay of peers would
unnecessarily increase the proportion of maintenance-related messaging, which is
usually dominant in DHT-based P2P networks. Because of its structural simplicity,
the PCMS can theoretically support an arbitrary combination of a DHT-based P2P
protocol and a DHT algorithm. This is enabled by two essential features of the
prototype: (1) the P2P protocol is not modified, but only used in a specific way to
enable unified access to the user communities, and (2) the community overlays
are independent from each other, and thus, different DHT algorithms can be used
for different community overlays.
The performance of the PCMS was experimentally evaluated in terms of
network traffic load, routing latencies, success ratio of look-up operations and
mobile nodes’ energy consumption. The PCMS was also compared against a flat
DHT structure where all the community-related activities are performed in a
single DHT overlay. In the evaluation, each node was run on the same dedicated
server machine as an independent instance of the PCMS prototype
implementation. The results related to the network traffic load indicate that the
PCMS performs especially well when the overall P2P network is large, the
community overlays are small and there is plenty of community-related activity
of the nodes. These advantages are mainly achieved because of more efficient
routing taking place in the smaller community overlays. The results related to
energy consumption indicate that with WLAN, the level of network activity and
the size of an average network packet had only a minor effect on the operational
time of a mobile device. With UMTS, however, the length of the operational time
was practically only dependent on the size of an average network packet. With
WLAN, the battery of a mobile device was consumed in six hours in average, and
with UMTS, only in two hours in average, which is obviously too limited use
period for today. In conclusion, the PCMS, as well as any other P2P community
management system in general could be further optimised for mobile devices by
adjusting the size of an average network packet. In this respect, more
56
experimental evaluation is needed using more diverse levels of network activity
and sizes of an average network packet.
Paper III, entitled “Robustness of a P2P Community Management System
based on Two-level Hierarchical DHT Overlays” continues the work started in
Paper II and evaluates the robustness of the PCMS in more detail. The
experimental evaluation was conducted using P2PP as the DHT-based P2P
protocol and both Kademlia and Chord as DHT algorithms. The robustness of the
PCMS was evaluated with varying levels of churn and node activity, in terms of
look-up success ratio, network traffic load and average hop count. The level of
churn was adjusted by changing the average node online time and the way the
nodes leave the P2P overlay network, which could occur either in a graceful or
ungraceful manner. Like in Paper II, the PCMS was compared against a flat DHT
structure where all the community related activities are performed in a single
DHT overlay. In the evaluation, each node was run on a dedicated server machine
as an independent instance of the PCMS prototype implementation.
To ensure fluent user experience of a community-centric service, the look-up
success ratio should not decrease below 50%. With graceful leaving of nodes, the
Kademlia-based PCMS had a very high look-up success ratio even with the
shortest node online times. The look-up success ratio of the Chord-based PCMS
was clearly lower, but still at an acceptable level. With ungraceful leaving of
nodes, the look-up success ratios decreased with both the Kademlia-based and the
Chord-based PCMS. With Kademlia-based PCMS, the look-up success ratio
stayed at an acceptable level, but with the Chord-based PCMS, the look-up
success ratio fell below 50% with the shortest node online times.
When compared against the flat DHT structure, the evaluation results indicate
that the PCMS enhances the performance of community management in a high
churn environment. With Kademlia, this is exemplified by a slight decrease in the
network traffic load and by a significant decrease in the hop count. With Chord, in
turn, this is exemplified by the significantly improved look-up success ratio with
node online time of 800 seconds or shorter, and also by a significant decrease in
the hop count. In the future, the performance of the PCMS could be improved by
fine-tuning the operation of the DHT algorithm and the DHT-based P2P protocol
to the deployment environment and/or to the specific application case. However,
these aspects have not been thoroughly studied in this thesis, and hence, further
research is required.
57
3.3 A novel community-centric mobile service environment
The third research question in this thesis concerned implementing the
communication infrastructure for community-centric mobile services using P2P
and Web technologies. Papers IV and V contain the contribution to this research
area and address the third research question.
Paper IV, entitled “Towards Context-aware Mobile Web 2.0 Service
Architecture” introduces initial considerations on the implementation of the
COMSE, and thus, provides a starting point for the research work done in this
thesis. The publication analyses different architectural approaches for
implementing a context-aware mobile Web 2.0 service architecture that connects
individual and community context information with Web-based services. In the
service architecture, four communication models are introduced for indexing and
discovery of context information and Web-based services. The communication
models are based on different degrees of centralisation.
The indexing and discovery of both context information and Web-based
services is implemented independently of each other using either CS or P2P
technologies. Based on the analysis, it is suggested that use of CS technologies is
a better choice for indexing and discovery of context information if availability
and freshness of context information (e.g. location) is crucial. This is because of
the fact that the search efficiency in P2P overlay networks cannot be usually
guaranteed. In this case, the load on the centralised servers should be alleviated
by using the P2P technologies for indexing and discovery of Web-based services.
Paper IV also illustrates the market structures of the proposed communication
models. The market structures are analysed from a multidisciplinary point of view
using the industrial organisation theory. According on the analysis, the
communication models that use P2P technologies are more probable to promote
perfect competition.
Paper V, entitled “Case Study on a Community-centric Mobile Service
Environment” continues the work started in Paper IV and introduces the COMSE,
a service environment for community-centric mobile services. The COMSE
advances the knowledge of using P2P and Web technologies collaboratively in
mobile environments by providing a general-purpose service environment, where
all functionalities of a P2P system can be seamlessly used by a Web-based service
running in a Web browser of a mobile device. In the COMSE, P2P technologies
enhance fairness of load distribution, scalability and fault tolerance, whereas the
Web-based services enable easy development, reuse and composition of platform
58
independent mobile services. The prototype of the COMSE was implemented for
N810 Internet tablet that is based on Linux Maemo. This particular mobile device
was selected as the development platform, because it was predicted to be a
trendsetter for the mobile devices in the future. The overall architecture of the
COMSE is divided into three layers, as illustrated in Fig. 9.
Fig. 9. A layered architectural model of the COMSE (Paper V).
On the protocol layer, reside P2PP with Kademlia and Chord DHTs, as well as
Sofia-SIP stack with Session Initiation Protocol for Instant Messaging and
Presence Leveraging Extensions (SIMPLE) (Rosenberg 2010). P2PP provides
functionality for managing P2P overlay networks and their resources, whereas the
Sofia-SIP provides functionality for encoding and decoding SIMPLE-specific SIP
messages that carry context information of individuals and user communities. On
the middleware layer, reside five management components, namely (1)
Community Management, (2) Community Access Management, (3) Content
Management, (4) Service Management and (5) Community Context Management,
of which each provides a particular functionality for the Web-based services on
the service layer: Community Management component provides functionality for
establishing communities using the PCMS presented in Paper II and Paper III;
Community Access Management component provides functionality for joining to
location-based communities using either RFID or BT technologies; Content
Management component provides functionality for using P2P overlay networks as
a distributed storage; Service Management provides functionality for indexing
and discovering Web-based services within the established communities; and
Contextservice
1) CommunityManagement
Web-based services
Service layer
Middleware layer
2) CommunityAccess
Management
3) ContentManagement
4) ServiceManagement
5) CommunityContext
Management
Sofia-SIP stackwith SIMPLE
P2PP with Kademlia and Chord DHT
Protocol layer
P2P Daemon
59
Community Context Management component provides functionality for storing
and delivering context information of individuals and user communities.
The management components are part of a mobile middleware called P2P
Daemon that acts as a gateway between the P2P overlay networks and the Web-
based services. P2P Daemon can communicate with one or multiple P2P overlay
networks by acting as a node in these P2P networks, whereas the communication
between the Web-based services and P2P Daemon is implemented using a specific
interface component called P2P Bridge. In the prototype of the COMSE, P2P
Bridge was implemented using an open source component called D-Bus that
enables communication between local applications on Linux-based platforms. In
this case, the local applications are the Web browser and P2P Daemon. Because
the interaction between P2P Daemon and the P2P overlay networks takes place
over a wireless network connection, there is, of course, always some network
delay present. Thus, all communication enabled by P2P Daemon was
implemented in an asynchronous manner. For the Web-based services, the
capability for asynchronous communication was implemented in P2P Bridge
using AJAX.
The performance of the COMSE was evaluated in terms of network traffic
load, success ratio of P2P operations, and latencies of successful P2P operations.
The evaluation took place in a real deployment environment with 53 test users.
The technical data needed for the performance evaluation was acquired
simultaneously while conducting the user studies of two pilot services (Papers VI
and VII) built on top of the COMSE. The two pilot services used the COMSE for
managing user communities, as well as for service indexing and discovery, but
not for storage. In the evaluation, each test user was equipped with a N810
Internet Tablet that was connected to the Internet using a WLAN base station
(IEEE 802.11b/g). Due to the capacity limitations of the available WLAN base
station, a P2P overlay network of approximately ten test users could be
established at a time. This resulted in five comparable test events. Although the
size of the established P2P overlay networks was small, a considerable amount of
churn was present, because some mobile nodes crashed several times. This was
caused by mistakes in the source code of the prototype, heavily loaded WLAN
base station, and the poorly performing WLAN connection of the N810 devices.
The performance evaluation results indicate that the average network traffic
load on a mobile node was not very high. According to the energy consumption
measurements conducted in Paper II, a mobile node could stay operational in this
kind of P2P overlay networks for up to eight hours. However, it should be taken
60
into account that the size of the established DHT-based P2P overlay network was
small, which resulted in a relatively small maintenance load. In DHT-based P2P
systems, the proportion of the maintenance load is usually dominating. The
overall success ratio of P2P operations was also high, regardless of the fact that
every third operation required one or more resubmissions. However, the look-up
success ratio was slightly above the appropriate level of 50% being more
significantly affected by the high level of churn. The latencies of successful P2P
operations were rather low as expected, because the mobile nodes were connected
to the same WLAN base station. Although the evaluation results support the
feasibility of the COMSE, more extensive experimental evaluation should be
conducted in the future. The COMSE or a similar kind of approach should be
tested with larger P2P overlay networks, and with community-centric mobile
services that use the P2P overlay networks in a more diverse manner.
3.4 User acceptance studies of community-centric mobile services
The fourth research question in this thesis deals with factors that affect user
acceptance of exemplary community-centric mobile services. Papers VI and VII
contain the contribution to this research area and address the fourth research
question. The CCMV and CNL services, evaluated in this section, have been built
on top of the COMSE that was presented in Paper V. Therefore, the CCMV and
the CNL services were also used as instruments for evaluating the feasibility of
the COMSE.
Paper VI, entitled “User Evaluation of a Community-centric Music Voting
Service” evaluates users’ perceptions and user acceptance of the CCMV service
that relies on the COMSE and a virtual disc jockey system called DjOnline. In
brief, DjOnline provides a large database of music enriched with metadata that
allows creating playlists automatically based on musical preferences. The music
played by DjOnline is streamed from a server to client front-ends that reside in
commercial premises. For the CCMV service, DjOnline was enhanced with the
functionality for creating music voting events.
In the CCMV service, a DHT-based P2P overlay network is established for
each location-based community. The joining parameters for a particular
community can be retrieved either from a BT beacon or an RFID tag installed in
the proximity of that community. The joining parameters contain, in practice, the
IP address and the port of one or multiple bootstrap nodes. Once a member of a
community, users are able to influence the music played within the premises by
61
participating in the music voting events. Users are also able to see the playlist of
the upcoming songs, to request songs to be played, and to interact with other
community members. In the prototype version of the CCMV, the interaction
included discovering the other community members, examining their personal
music profiles and sending messages to a common messaging board.
Fig. 10. The Decomposed Theory of Planned Behaviour (Paper VI).
The user study of the CCMV service was conducted in a real deployment
environment with 53 test users who represented the clientele of a bar. The
evaluation took place in the city of Oulu, Finland. The factors affecting user
acceptance of the CCMV service analysed using a research model called DTPB.
The version of DTPB, defined for studying community-centric mobile services, is
illustrated in Fig. 10. In Fig. 10, the significance of each determinant on the use
intention towards the CCMV service is shown. An asterisk (*) denotes that the
hypothesis was not supported. The analysis of users’ perceptions on the CCMV
service was based on the outcome of short group interviews and the differences in
Attitude
Subjectivenorm
Behaviouralintention
Perceivedbehavioural
control
Internetexperience
Mobileexperience 0.41
0.19*
0.41
Friendinfluence
Culturalinfluence
0.18*
0.42
0.49
Perceivedplayfulness
Perceivedusefulness
Image
Perceivedprice level -0.18*
0.45
0.22
0.16
0.02*
62
the perceptions between the defined three special interest groups, namely people
who are very enthusiastic about new technology, people who are very enthusiastic
about listening to music, and people who use their mobile phone for listening to
music.
The evaluation results indicate that cultural-related habits and norms, as well
as earlier experiences with mobile devices, are the two most important factors that
affect the use intentions towards the CCMV service. The test users brought forth
that using the CCMV service should be very straight-forward and should not
interfere with the customary behaviour, for instance, in a bar. In general, the
voting for music in a community was experienced to be great fun. However, not
all the test users were convinced that the music voting in a community should be
arranged as a continuous activity, but rather as a scheduled wish hour. This was a
result from the following two facts: the concern that being in a bar would turn
into tapping out a mobile device, and the preference to hear music played also by
a professional disc jockey. In the midst of music voting, some users also preferred
a chat service that could be used for approaching new interesting people. Related
to this, adding a symbol that describes whether a person is a single or not was
requested. At the same time, however, several users voiced a privacy concern
about the storage and the visibility of users’ personal data.
Test users also clearly pointed out that use of the CCMV service should be
free-of-charge. However, they did not oppose adding the service expenses to the
prices of other services provided in the entertainment premises. These findings
should be taken into account when defining the business model for the CCMV
service. For instance, advertisement-based or revenue-sharing business models
could be applied. To implement revenue-sharing, the CCMV service could be
interfaced with a mobile music store. As a result, any song played within the
entertainment premises could be found and purchased from the mobile music
store. With each successful purchase, part of the revenue would then be shared
with the owner of the entertainment premises.
Based on the analysis of the defined interest groups, test users who were very
keen on technology, listening to music and using a mobile device as a music
consumption medium were found to be potential early adopters of the CCMV
service. These test users expressed stronger intention to use the CCMV service
than an average test user.
Paper VII, entitled “User Experience in Added Value Location-based Mobile
Music Service” evaluates users’ perceptions on the CNL service that relies on the
COMSE and DjOnline. The idea behind the CNL service is to provide
63
recommendations about entertainment premises in the proximity of a user based
on her music preferences. The music profiles of entertainment premises are
retrieved from DjOnline and matched with the user’s personal music profile. The
matching of the music profiles was originally conducted using an algorithm
presented in Paper VII, but later by calculating multidimensional Euclidean
distance. The entertainment premises are displayed on a map-based user interface
and are coloured grey, yellow or green according to their match.
The functionality of the CNL service can also be extended. This is achieved
by adding light-weight Web-based services such as Pictures Nearby (Järvinen et
al. 2009b) to the map-based user interface. By implementing the application
programming interface (API) provided by the CNL service, the integration is
seamless from the user’s viewpoint. For indexing and discovery of these services,
a DHT-based P2P overlay network is established. All in all, the CNL service
demonstrates how to use Web as a service platform in a mobile environment.
The user study of the CNL service was conducted in the same environment
and with the same test users as the user study of the CCMV service. The analysis
of users’ perceptions on the CNL service was based on the outcome of short group
interviews and on the differences in the perceptions between the same three
interest groups created according to test users’ technological awareness and music
consumption habits. The evaluation results indicate that the CNL service would
be the most beneficial in relatively large and unfamiliar cities. However, it was
also emphasised several times that it should be possible to define one’s own
friend communities and to find out the bars one’s friends are visiting at the
moment. This finding suggests that an extended version of the CNL would also be
useful in a more familiar neighbourhood. The evaluation results further indicate
that the options for defining the personal music profile were perceived to be too
abridged. This was partly expected, because the music profile copied from
DjOnline was simplified notably. Based on this finding, it appears that the more
diverse music profile would probably be created already at home and only fine-
tuned while on the move. In addition, the test users requested that the personal
music profile could be created semi-automatically based on their favourite songs.
From the three interest groups, test users who regularly used their mobile devices
for listening to music were clearly the most eager to start using the CNL service.
64
3.5 Discussion and future work
With the rapidly maturing mobile technologies, novel community-centric mobile
services that are established according to the shared location and/or activities of
people are becoming technically feasible. This thesis casts light on the
development of these novel mobile services by providing new knowledge and a
source of future ideas for the developers and the researchers of P2P systems,
mobile services, as well as mobile middleware.
The conceptualisation and the survey of P2P GMSs, presented in Paper I, are
pioneering attempts to provide a comprehensive understanding of P2P group
management. Based on the knowledge provided in Paper I, the selection of a
suitable P2P system architecture for implementing a particular kind of P2P GMS
can be facilitated. Although many P2P systems that are composed of multiple
connected overlays have been already proposed in the literature (e.g. Xu et al.
2003, Artigas et al. 2005), they have not been experimented for managing
location and/or activity-based user communities and have not been evaluated
using real-life prototype implementations. The latter is an important aspect,
because according to Kato & Kamiya (2007), the results gained using simulations
may differ significantly from the measurements conducted with real-life
prototype implementations. The PCMS that was experimentally evaluated in
Papers II and III provides knowledge how a P2P system that is composed of
multiple connected DHT-based overlays suits for managing user communities.
Despite the fact that the use of DHT-based P2P systems in dynamic mobile
environments has been questioned by the researchers of P2P systems, the results
in Papers II, III and V indicate that Kademlia DHT carries off well in a dynamic
environment when P2P overlay networks are relatively small-sized and the
average node online time is not extremely short.
The integrated use of P2P and Web technologies in mobile environments is
becoming a popular research topic as witnessed by the increasing number of
research papers. The COMSE, introduced in Paper V, provides a general-purpose
service environment, where all functionalities of a P2P system can be directly
used by a Web-based service running in a Web browser of a mobile device.
According to the author’s best knowledge, this approach is novel, because the
existing works such as MyNet (Kalofonos et al. 2008) use P2P technologies
merely for implementing a GMS and the service discovery of Web-based services,
but not for enabling interaction between a P2P system and a Web-based service.
In the COMSE, a Web-based service can directly access any available P2P-based
65
functionality. Currently, the COMSE is limited to Linux-based environments,
because the prototype version of the COMSE was developed for Linux Maemo-
based N810 Internet tablet.
Although some user acceptance studies on community-centric services exist
(e.g. Van Slyke et al. 2002, Dholakia et al. 2004), their focus is primarily on the
fixed environment. Papers VI and VII introduce two exemplary community-
centric mobile services, namely CCMV and CNL, and provide some insight into
the aspects that need to be taken into account when designing and deploying
novel community-centric mobile services. The results indicate that the most
important aspects to consider are the fit between a community-centric mobile
service and its intended usage situation, as well as the target user group’s level of
experience in mobile devices. The first aspect has also been discovered, for
instance, by Slyke et al. (2002) in their study of the intentions to use a groupware.
Because the implementations of the CCMV and the CNL services were merely
prototypes, some of the intended community-centric features could not be
implemented. This was a limitation that complicated a comprehensive testing of
community-centric aspects of the CCMV and the CNL services. Furthermore, the
test period of 45 minutes was rather short in order to enable a thorough
examination of the two exemplary community-centric mobile services. However,
as pointed out by Koivumäki et al. (2008), the first impression is especially
important with mobile services. According to their study, users’ perceptions did
not differ significantly depending on the time spent on using the examined mobile
services. Finally, it is worth to mention that some features of the CCMV service
have also been applied in a commercially deployed DjOnline service.
Despite the results achieved in this thesis, there remain several research areas
on the community-centric mobile services that should be examined further. Based
on the battery life measurements, conducted in Paper II, it was discovered that the
size of an average network package has a notable effect on the battery life of a
mobile device. Thus, the battery life of a mobile device could be extended by
optimising the size of an average network packet. In this respect, more
experimental evaluation is needed using more diverse levels of network activity
and sizes of an average network packet. Furthermore, the performance of any
DHT-based P2P community management system could be improved by fine-
tuning the operation of the selected DHT algorithm and the DHT-based P2P
protocol according to the deployment environment and/or to the specific
application case. Hautakorpi & Camarillo (2007) have already studied the
suitability of different DHT algorithms from the standpoint of interpersonal
66
communications, but the same examination should also be conducted with various
types of community-centric mobile services.
Although the evaluation results, presented in Paper V, support the feasibility
of the COMSE, more extensive experimental evaluation should be conducted in
the future. The COMSE or a similar kind of approach should be tested with larger
P2P overlay networks, and with community-centric mobile services that use the
P2P overlay networks in a more diverse manner. Furthermore, more user studies
focusing on both user acceptance and UX in community-centric mobile services
are required in order to increase the understanding of the critical success factors.
A longitudinal user study would also reveal whether the technical optimisations
that strive to preserve the battery life of a mobile device have a significant effect
on either user acceptance or UX in community-centric mobile services.
For increasing the applicability of community-centric mobile services for
varying environments, the possibilities of using CS, P2P and MANET
technologies collaboratively should be examined more thoroughly in the future.
This also includes considerations for taking advantage of the information that
originates from the established SNSs. For instance, the community-centric mobile
services could be enriched with the existing information on personal profiles,
specific entertainment events and community memberships.
67
4 Conclusions
The revolution of social networking started a couple of years after the turn of the
millennium, when several popular SNSs were introduced. The importance of
social networking is obvious, because today, even mobile devices may have
buttons and widgets that are dedicated for popular SNSs. As the next evolutionary
step for social networking, user communities will be dynamically established
according to contextual aspects such as the shared location and/or activities of
people. Taking this step is primarily enabled by the rapid development of mobile
devices that are now equipped with decent processing capabilities, multiple
sensors and several wireless network interfaces. In the development of the
community-centric mobile services, P2P technologies have their implicit
advantages over the dominating CS technologies. However, the two technologies
are not mutually exclusive, but can also be used efficiently in collaboration.
In this thesis, an experimental prototype of the COMSE was developed. All
services provided in the COMSE were implemented using platform-independent
Web technologies. The central component of the COMSE is a mobile middleware
called P2P Daemon that enables the use of P2P overlay networks and context
information as part of Web-based services running on a Web browser of a mobile
device. In the COMSE, P2P overlay networks are primarily used for enabling user
communities. To address this, the advantages and disadvantages of different P2P
system architectures were studied for implementing the management of
communities. As a result, an experimental prototype of the PCMS was developed
and its performance was evaluated. In the PCMS, each community is
implemented as an independent DHT-based P2P overlay network that are
advertised in a global DHT-based P2P overlay network. The experimental
prototype of the COMSE was used for two purposes: first, for studying the
feasibility of using P2P and Web technologies in mobile environments, and
second, for studying users’ perceptions and user acceptance of two exemplary
community-centric mobile services, namely CCMV and CNL. The evaluation was
conducted in a real deployment environment with 53 test users. In the user studies,
the main aim was to identify the factors that affect user acceptance of new
community-centric mobile services.
The development of new mobile services is a challenging task. This is a result
from the fact that users’ perceptions on new mobile services are very hard to
predict. The community-centric mobile services, introduced in this thesis, were
not only new types of mobile services to the users, but also to the developers of
68
the service prototypes. This made it especially difficult to anticipate the users’
perceptions before carrying out the testing of the service prototypes in a real-life
usage environment.
While conducting the research work presented in this thesis, it became
evident that it is crucial to understand the motivating factors that guide users’
activity in the real-life usage environment. In this context, it is also essential to
find out how the local culture, generally accepted behaviour in particular,
influences users’ interaction with novel community-centric mobile services. For
instance, some test users voiced that they would consider it inappropriate if most
of the people were frequently tapping out their mobile devices in entertainment
premises such as bars. These findings emphasise the importance of involving the
users in the software development process, already in the very early phases, to
understand all the facets of the requirements for a community-centric mobile
service. With the two exemplary community-centric mobile services, introduced
in this thesis, the requirement of having an intuitive user interface that contains
only the necessary functionality was emphasised. In addition, several users voiced
their concerns about the security of their personal information. This probably
stems from the facts that a mobile device is considered a more personal device
than a desktop computer at home (typically used for SNSs), and that P2P as a
technology is not generally well-known and trusted. It also seems that P2P still
partly carries the stigma of being illegal though the blame should not be on the
technology itself, but on the people who use it for wrong purposes.
Due to the rapid development of mobile technologies, enterprises are not
likely aware of what is now technically viable. As a consequence, the enterprises
may not realise what kind of new mobile services they could, possibly in
collaboration with the other players, provide to their customers. In this context,
also the mobile operators should be encouraged to see the potential of mobile P2P
technologies, and not only the anticipated threat of excessive bandwidth usage. To
achieve this, successful deployments of community-centric mobile services are
required. This would provide evidence that the introduction of community-centric
mobile services may result in an appreciable growth of the mobile service market.
From the academic perspective, the development of community-centric
mobile services can be facilitated by creating new knowledge on the feasibility of
the alternative technologies, on the factors that affect user acceptance, and also,
on the viability of different business models. In order to have a comprehensive
understanding, the three tasks should not be conducted separately, but in close
collaboration. For this, a multidisciplinary research approach must be adopted.
69
References
Adibi S & Erfani S (2006) A multipath routing survey for mobile ad-hoc networks. Proc IEEE Consumer Communications and Networking Conference. Las Vegas, USA: 984–988.
Agarwal R & Pradat J (1997) The role of innovation characteristics and perceived voluntariness in the acceptance of information technologies. Decision Sciences 28(3): 557–582.
Agostini A & Moro G (2004) Identification of communities of peers by trust and reputation. Lecture Notes in Computer Science: Artificial Intelligence - Methodology, Systems and Applications 319: 85–95.
Ajjan H & Hartshorne R (2009) Investigating faculty decisions to adopt Web 2.0 technologies: theory and empirical tests. The Internet and Higher Education 11(2): 71–80.
Ajzen I (1991) The theory of planned behaviour. Organisational behaviour and human decision processes 50(2): 179–211.
Androutsellis-Theotokis S & Spinellis D (2004) A survey of peer-to-peer content distribution technologies. ACM Computing Surveys 36(4): 335–371.
Artigas M, López P & Skarmeta A (2005) Cyclone: a novel design schema for hierarchical DHTs. Proc IEEE International Conference on Peer-to-Peer Computing. Konstanz, Germany: 49–56.
Baset S & Schulzrinne H (2004) An analysis of the Skype peer-to-peer Internet telephony protocol. Technical Report CUCS-039-04. New York, USA: 1–12.
Baset S, Schulzrinne H & Matuszewski M (2007) Peer-to-Peer Protocol (P2PP). IETF. URI: http://tools.ietf.org/html/draft-baset-p2psip-p2pp-01. (work-in-progress).
Bharambe AR, Agrawal M & Seshan S (2004) Mercury: supporting scalable multi-attribute range queries. Proc ACM Conference on Applications, Technologies, Architectures, and Protocols for Computer Communications. Portland, USA: 353–366.
Birman K, Cantwell J, Freedman D, Huang Q, Nikolov P & Ostrowski K (2009) Building collaboration applications that mix Web Services hosted content with P2P protocols. Proc IEEE International Conference on Web Services. Los Angeles, USA: 509–518.
Bottazzi D, Montanari R & Rossi G (2008) A self-organising group management middleware for mobile ad-hoc networks. Computer Communications 31: 3040–3048.
Boyd DM & Ellison NB (2007) Social network sites: definition, history and scholarship. Journal of Computer-Mediated Communication 13(1): 1–18.
Brands EHTB & Kragiannis G (2009) Taxonomy of P2P applications. Proc IEEE Workshop on Enabling Service-Oriented Internet. Hawaii, USA: 1–8.
Capra L, Emmerich W & Mascolo C (2003) CARISMA: context-aware reflective middleware system for mobile applications. IEEE Transactions on Software Engineering 29(10): 929–944..
Castano S & Montanelli S (2005) Semantic self-formation of communities of peers. Proc ESWC Workshop on Ontologies in Peer-to-Peer Communities. Heraklion, Crete: 1–12.
70
Castro M, Druschel P, Kermarrec A-M & Rowstron A (2002) SCRIBE: a large-scale and decentralised application-level multicast infrastructure. IEEE Journal on Selected Areas in Communications 20(8): 100–110.
Chan ATS & Chuang S-N (2003) MobiPADS: a reflective middleware for context-aware mobile computing. IEEE Transactions on Software Engineering 29(12): 1072–1085.
Chen G & Kotz D (2000) A survey of context-aware mobile computing research. Technical Report TR2000-381. Hanover, USA: 1–16.
Cheong JH & Park M-C (2005) Mobile internet acceptance in Korea. Internet Research 15(2): 125–140.
Chlamtac I, Conti M & Liu JJN (2003) Mobile ad hoc networking: imperatives and challenges. Ad Hoc Networks 1(1): 13–64.
Constantine LL & Lockwood LAD (2002) User-centred engineering for web applications. IEEE Software 19(2): 42–50.
Das T, Nandi S & Ganguly N (2009) Community formation and search in P2P: a robust and self-adjusting algorithm. Proc Workshop on Intelligent Networks: Adaptation, Communication & Reconfiguration. Bangalore, India: 1–8.
Davis FD, Bagozzi RP & Warshaw PR (1989) User acceptance of computer technology: a comparison of two theoretical models. Management Science 35(8): 982–1003.
Deshpande M, Amit A, Chang M, Venkatasubramanian N & Mehrotra S (2007) Flashback: a peer-to-peer Web server for flash crowds. Proc International Conference on Distributed Computing Systems. Toronto, Canada: 1–8.
Dey A, Abowd G & Salber D (2001) A conceptual framework and a toolkit for supporting the rapid prototyping of context-aware applications. A Special Issue on Context-Aware Computing in Human-Computer Interaction Journal 16(2–4): 97–166.
Dey A, Hightower J, de Lara E & Davies N (2010) Location-based services. IEEE Pervasive Computing 9(1): 11–12.
Dillon A (2001) User acceptance of information technology. In: Karwowski W (ed) Encyclopedia of Human Factors and Ergonomics. London, Taylor and Francis.
Dix A, Finlay J, Abowd GD & Beale R (2004) Human-computer interaction (3rd edition). Harlow, England. Pearson Education Limited.
Dholakia UM, Bagozzi RP & Pearo LK (2004) A social influence model of consumer participation in network- and small-group-based virtual communities. International Journal of Research in Marketing 21(3): 241–263.
Eckhardt A, Laumer S & Weitzel T (2009) Who influences whom? Analysing workplace referents’ social influence on IT adoption and non-adoption. Journal of Information Technology 24(1): 11–24.
Facebook (2011) Press Room. URI: http://www.facebook.com/press. Ferscha A, Holzmann C & Oppl S (2004) Context-awareness for group interaction support.
Proc International Workshop on Mobility Management & Wireless Access Protocols. Philadelphia, USA: 88–97.
Fishbein M & Ajzen I (1975) Belief, attitude, intention, and behaviour: an introduction to theory and research. Reading, MA, USA, Addison-Wesley.
71
Ford B, Strauss J, Lesniewski-Laas C, Rhea S, Kaashoek F & Morris R (2006) Persistent personal names for globally connected mobile devices. Proc USENIX Symposium on Operating Systems Design and Implementation. Seattle, USA: 1–16.
Forlizzi J & Battarbee K (2004) Understanding experience in interactive systems. Proc Conference on Designing Interactive Systems: Processes, Practices, Methods, and Techniques. Cambridge, Massachusetts, USA: 261–268.
Fouskas KG, Giaglis GM, Kourouthanassis PE, Karnouskos S, Pitsillides A & Stylianou M (2005) A roadmap for research in mobile business. International Journal of Mobile Communications 3(4): 350–373.
Gehlen G & Pham L (2005) Mobile Web Services for peer-to-peer applications. Proc IEEE Consumer Communications and Networking Conference. Las Vegas, USA: 427–433.
Gong L (2001) JXTA: a network programming environment. IEEE Internet Computing 5(3): 88–95.
Gu T, Pung HK & Zhang DQ (2004) A middleware for building context-aware mobile services. Proc Vehicular Technology Conference. Milan, Italy: 2656–2660.
Gulliksen J, Göransson B, Boivie I, Blomkvist S, Persson J & Cajander Å (2003) Key principles for user-centred systems design. Behaviour and Information Technology 22(6): 397–409.
Gupta S & Kim H-W (2004) Virtual community: concepts, implications, and future research directions. Proc Americas Conference on Information Systems. New York, USA: 2679–2687.
Harjula E, Ylianttila M, Ala-Kurikka J, Riekki J & Sauvola J (2004) Plug-and-play application platform: towards mobile peer-to-peer. Proc International Conference on Mobile and Ubiquitous Multimedia. College Park, USA: 63–69.
Hassan M (2009) Mobile Web Service provisioning in peer-to-peer environments. Proc IEEE International Conference on Service-Oriented Computing and Applications. Taipei, Taiwan: 1–4.
Hassenzahl M & Tractinsky N (2006) User experience – a research agenda. Behaviour and Information Technology 25(2): 91–97.
Hautakorpi J & Camarillo G (2007) Evaluation of DHTs from the viewpoint of interpersonal communications. Proc International Conference on Mobile and Ubiquitous Multimedia. Oulu, Finland: 74–83.
Haverinen H, Siren J & Eronen P (2007) Energy consumption of always-on application WCDMA networks. Proc Semi-Annual Vehicular Technology Conference. Dublin, Ireland: 964–968.
Henricksen K & Indulska J (2006) Developing context-aware pervasive computing applications: models and approach. Pervasive and Mobile Computing 2(1): 37–64.
Hightower J, Brumitt B & Borriello G (2002) The location stack: a layered model for location in ubiquitous computing. Proc IEEE Workshop on Mobile Computing Systems and Applications. New York, USA: 1–7.
International Data Corporation (2011) IDC Forecasts Worldwide Smartphone Market to Grow by Nearly 50% in 2011. Press Release. URI: http://www.idc.com/ getdoc.jsp?containerId=prUS22762811.
72
International Organization for Standardization (2009) Ergonomics of human system interaction - Part 210: Human-centred design for interactive systems. ISO 9241-210.
International Telecommunication Union (2009) Overall network operation, telephone service, service operation and human factors. Recommendation ITU-T E.800.
Iwamaru A, Itokawa T, Kitasuka T & Aritsugi M (2009) Introducing group participation support into P2P Web caching systems. Proc International Conference on Advanced Information Networking and Applications. Bradfod, UK: 868–875.
Järvinen S, Peltola J, Lahti J & Sachinopoulou A (2009a) Multimedia service creation platform for mobile experience sharing. Proc International Conference on Mobile and Ubiquitous Multimedia. Cambridge, UK: 1–9.
Järvinen S, Peltola J, Plomp J, Ojutkangas O, Heino I, Lahti J & Heinilä J (2009b) Deploying mobile multimedia services for everyday experience sharing. Proc IEEE International Conference on Multimedia Expo. New York, USA: 1760–1763.
Johnson RR, Salvo MJ & Zoetewey MW (2007) User-centred technology in participatory culture: two decades “beyond a narrow conception of usability testing”. IEEE Transactions on Professional Communication 50(4): 320–332.
Junglas IA & Watson RT (2008) Location-based services: evaluating user perceptions of location-tracking and location-awareness services. Communications of the ACM 51(3): 65–69.
Kalofonos DN, Antoniou Z, Reynolds FD, Van-Kleek M, Strauss J & Wisner P (2008) MyNet: a platform for secure P2P personal and social networking services. Proc IEEE International Conference on Pervasive Computing and Communications. Hong Kong, China: 135–146.
Kato D & Kamiya T (2007) Evaluating DHT implementations in complex environments by network emulator. Proc International Workshop on Peer-to-Peer Systems. Bellewue, USA.
Kelényi I & Nurminen JK (2008) Optimizing energy consumption of mobile nodes in heterogeneous Kademlia-based distributed hash tables. Proc Conference on Next Generation Mobile Applications, Services, and Technologies. Cardiff, UK: 70–75.
Khabatti, M, Ryu K & Dasgupta P (2002) Peer-to-peer communities: formation and discovery. Proc International Conference on Parallel and Distributed Computing Systems. Cambridge, USA: 161–166.
Kilkki K (2008) Quality of experience in communications ecosystem. Journal of Universal Computer Science 14(5): 615–624.
Koivumäki T, Ristola A & Kesti M (2008) The perceptions towards mobile services: an empirical analysis of the role of use facilitators. Personal and Ubiquitous Computing 12(1): 67–75.
Kortuem G (2002) Proem: a middleware platform for mobile peer-to-peer computing. Mobile Computing and Communications Review 6(4): 62–64.
Koskela T, Julkunen J, Keränen V, Kostamo N & Ylianttila M (2009) User experiences on a community-based music voting service. Proc 3rd International Workshop on Service Intelligence and Computing. Los Angeles, USA.
73
Kozierok CM (2005) The TCP/IP guide: a comprehensive, illustrated Internet protocols reference. San Francisco, USA. No Starch Press.
Krishnamurthy B, Wang J, & Xie Y (2001) Early measurement of a cluster-based architecture for P2P systems. Proc ACM SIGCOMM Internet Measurement Workshop. San Francisco, USA: 105–109.
Krogstie J, Lyytinen K, Opdahl A, Pernici B, Siau K & Smolander K (2003) Mobile information systems - research challenges on the conceptual and logical level. In: Olivé A (ed) Advanced Conceptual Modeling Techniques. Lecture Notes in Computer Science 2784: 124–135.
Kushniruk AW & Patel VL (2004) Cognitive and usability engineering methods for the evaluation of clinical information systems. Journal of Biomedical Informatics 37(1): 56–76.
Lechner U & Hummel J (2002) Business models and system architectures of virtual communities: from a sociological phenomenon to peer-to-peer architectures. International Journal of Electronic Commerce 6(3): 41–53.
Lee IW, Choi M-O, Park H-J & Park K-R (2008) Social networking service based on peer-to-peer network. Proc International Conference on Systems and Networks Communications. Sliema, Malta: 235–238.
Lee WJ, Kim TU & Chung J-Y (2002) User acceptance of the mobile internet. Proc Conference on Mobile Business. Athens, Greece: 1–9.
Legrisa P, Inghamb J & Collerettec P (2003) Why do people use information technology? A critical review of the technology acceptance model. Information and Management 40(3): 191–204.
Leibnitz K, Hoßfeld T, Wakamiya N & Murata M (2007) Peer-to-peer vs. client/server: reliability and efficiency of a content distribution service. Proc International Teletraffic Congress. Ottawa, Canada: 1–13.
Leonard D, Yao Z, Wang X & Loguinov D (2008) On Static and Dynamic Partitioning Behavior of Large-Scale P2P Networks. IEEE/ACM Transactions on Networking 16(6): 1475–1488.
Li J & Vuong S (2006) Grid resource discovery based on semantic P2P communities. Proc ACM Symposium on Applied Computing. Dijon, France: 754–758.
Li R, Zhang Z, Wang Z, Song W & Lu Z (2005). WebPeer: a P2P-based system for publishing and discovering Web Services. Proc. IEEE Services Computing Conference. Orlando, USA: 149–156.
Lu Y, Zhou T, & Wang B (2009) Exploring Chinese users’ acceptance of instant messaging using the theory of planned behavior, the technology acceptance model, and the flow theory. Computers in Human Behaviour 25(1): 29–39.
Lua EK, Crowcroft J, Pias M, Sharma R & Lim S (2005) A survey and comparison of peer-to-peer overlay network schemes. IEEE Communications Surveys 7(2): 72–93.
Maymounkov P & Mazieres D (2002) Kademlia: a peer-to-peer information system based on the XOR metric. Proc International Workshop on Peer-to-Peer Systems. Cambridge, Massachusetts, USA: 1–6.
74
Meier R & Cahill V (2010) On event-based middleware for location-aware mobile applications. IEEE Transactions on Software Engineering 36(3): 409–430.
Moore GC & Benbasat I (1991) Development of an instrument to measure the perceptions of adopting an information technology innovation. Information Systems Research 2(3): 192–222.
Morr CE & Kawash J (2007) Mobile virtual communities research: a synthesis of current trends and a look at future perspectives. International Journal of Web Based Communities 3(4): 386–403.
Muhammad N, Chiavelli D, Soldani D & Li M (2006) Introduction. In: Soldani D, Li M & Cuny R (ed) QoS and QoE management in UMTS cellular systems. Chichester, John Wiley & Sons: 1–8.
Nguyen T, Loke SW, Torabi T & Lu H (2008) Multiagent place-based virtual communities for pervasive computing. IEEE Workshop on Agent Technologies for Pervasive Communities. Hong Kong, China: 602–608.
Nielsen J (1993) Usability engineering. San Francisco, USA, Morgan Kaufmann. Norman D & Draper S (1986) User-centred system design: new perspectives on human-
computer interaction. Hillsdale, NJ, Lawrence Earlbaum Associates. Noury R, Fleury A, Rumeau P, Bourke AK, Laighin GÓ, Rialle V & Lundy JE (2007) Fall
detection - principles and methods. Proc International Conference of the IEEE Engineering in Medicine and Biology Society. Lyon, France: 1663–1666.
Ohtonen J, Kassinen O & Ylianttila M (2006) Feasibility study of a mobile peer-to-peer navigation application. Proc International Symposium on Personal, Indoor and Mobile Radio Communications. Helsinki, Finland: 1–5.
Ou Z, Harjula E, Kassinen O & Ylianttila M (2010) Performance evaluation of a Kademlia-based communication-oriented P2P system under churn. Elsevier Journal of Computer Networks 54(5): 689–705.
Pagani M (2004) Determinants of adoption of third generation mobile multimedia services. Journal of Interactive marketing 18(3): 46–59.
Paulson LD (2005) Building rich Web applications with Ajax. Computer 38(10): 14–17. Porter E (2004) A typology of virtual communities: a multi-disciplinary foundation for
future research. Journal of Computer-Mediated Communication 10(1). Prescott M & Conger S (1995) Information technology innovations: A classification by IT
focus of impact and research approach. ACM SIGMIS Database 26(2–3): 20–41. Ratnasamy S, Francis P, Handley M, Karp R & Shenker S (2001) A scalable content-
addressable network. Proc Conference on Applications, Technologies, Architectures and Protocols for Computer Communications. San Diego, USA: 161–172.
Reichl P (2007) From ‘Quality-of-Service’ and ‘Quality-of Design’ to ‘Quality-of-Experience’: a holistic view on future interactive telecommunication services. Proc International Conference on Software, Telecommunications and Computer Networks. Split-Dubrovnik, Croatia: 1–6.
Rheingold H (2000) The virtual community: homesteading on the electronic frontier. London, UK. MIT Press.
Rogers E (1962) Diffusion of innovations. New York, USA. The Free Press.
75
Rogers E (2003) Diffusion of innovations (5th edition). New York, USA. The Free Press. Rosenbaum S, Rohn JA & Humburg J (2000) A toolkit for strategic usability: results from
workshops, panels and surveys. In: Turner T, Szwillius G, Czerwinski M & Paterno F (ed) CHI 2000 Conference on Human Factors in Computing Systems Proceedings. ACM Press: 1–6.
Rosenberg J (2010) SIMPLE made simple: an overview of the IETF specifications for instant messaging and presence using the session initiation protocol (SIP). IETF draft. URI: http://tools.ietf.org/html/draft-ietf-simple-simple-06. (work-in-progress).
Rowstron A & Druschel P (2001) Pastry: scalable, distributed object location and routing for large-scale peer-to-peer systems. Proc IFIP/ACM International Conference on Distributed Systems Platforms. Heidelberg, Germany: 329–350.
Schroeder T, Goddard S & Ramamurthy B (2000) Scalable Web server clustering techniques. IEEE Network 14(3): 38–45.
Schubert P & Hampe JF (2005) Business models for mobile communities. Proc Hawaii International Conference on System Sciences. Hawaii, USA: 1–11.
Sheppard BH, Hartwick J & Warshaw PR (1988) The theory of reasoned action: A meta-analysis of past research with recommendations for modifications and future research. Journal of Consumer Research 15(3): 325–343.
Shin D-H (2007) User acceptance of mobile Internet: Implication for convergence technologies. Interacting with Computers 19(4): 472–483.
Srirama SN & Prinz W (2007) Mobile Web Services mediation framework. Proc Workshop on Middleware for Service Oriented Computing. Newport Beach, California, USA: 6–11.
Srirama SN, Jarke M & Prinz W (2006) Mobile Web service provisioning. Proc Advanced International Conference on Telecommunications and International Conference on Internet and Web Applications and Services. Guadeloupe, French Caribbean: 1–6.
Stoica I, Morris R, Karger D, Kaashoek M & Balakrishnan H (2001) Chord: a scalable peer-to-peer lookup service for Internet applications. Proc Conference on Applications, Technologies, Architectures, and Protocols for Computer Communications. San Diego, USA: 149–160.
Strang T & Linnhoff-Popien C (2004) A context modelling survey. Proc Workshop on Advanced Context Modelling, Reasoning and Management. Nottingham, UK: 1–8.
Sugano H, Fujimoto S, Klyne G, Bateman A, Carr W & Peterson J (2004) Presence information data format (PIDF). IETF RFC3863. URI: http://tools.ietf.org/html/rfc3863.
Tallberg M, Hämmäinen H, Töyli J, Kamppari S & Kivi A (2007) Impacts of handset bundling on mobile data usage: the case of Finland. Telecommunications Policy 31: 648–659.
Tarkoma S (2009) Mobile middleware architecture, patterns and practice. West Sussex, UK. John Wiley & Sons Ltd.
Tasch A & Brakel O (2004) Location based community services: new services for a new type of Web communities. Proc IADIS Conference on Web Based Communities. Lisbon, Portugal: 1–9.
76
Taylor S & Todd PA (1995) Understanding information technology usage: a test of competing models. Information Systems Research 6(2): 144–176.
Teo TSH & Pok SH (2003) Adoption of WAP-enabled mobile phones among Internet users. Omega 31(6): 483–498.
Terrace J, Laidlaw H, Liu HE, Stern S & Freedman MJ (2009) Bringing P2P to the Web: security and privacy in the Firecoral network. Proc International Conference on Peer-to-Peer Systems. Boston, USA: 1–7.
Tornatzky LG & Klein KJ (1982) Innovation characteristics and innovation adoption-implementation: a meta-analysis of findings. IEEE Transactions on Engineering Management 29(1): 28–45.
Urnadeta G, Pierre G & Van Steen M (2011) A survey of DHT security techniques. ACM Computing Surveys 43(2): 1–49.
Van Slyke C, Lou H & Day J (2002) The impact of perceived innovation characteristics on intentions to use groupware. Information Resources Management Journal 15(1): 5–12.
Venkatesh V & Davis FD (2000) A theoretical extension of the technology acceptance model: four longitudinal field studies. Management Science 46(2): 186–204.
Venkatesh V, Morris MG, Davis GB & Davis FD (2003) User acceptance of information technology: toward a unified view. MIS Quarterly 27(3): 425–478.
Väänänen-Vainio-Mattila K, Saarinen P, Wäljas M, Hännikäinen M, Orsila H, Kiukkonen N (2010) User experience of social ad hoc networking: findings from a large-scale field trial of TWIN. Proc ACM Conference on Mobile and Ubiquitous Multimedia. Limassol, Cyprus: 1–10.
Wanlei Z & Casey J (2008). Reducing cache lookup and publish overhead through locality awareness in P2P Web caching. Proc IEEE International Conference on High Performance Computing and Communications. Dalian, China: 296–303.
Winterbottom J, Thomson M & Tschofenig H (2009) GEOPRIV presence information data format location object (PIDF-LO) - usage clarification, considerations, and recommendations. IETF RFC5491. URI: http://tools.ietf.org/html/rfc5491.
Xiao Y, Tao Y, Li W & Li Q (2008) An effective P2P collaborative mode for mobile Web surfing. Proc International Conference on Network and Parallel Computing. Shanghai, China: 235–241.
Xu Z, Min R & Hu Y (2003) HIERAS: a DHT based hierarchical P2P routing algorithm. Proc International Conference on Parallel Processing. Kaohsiung, Taiwan: 187–194.
Yang B & Garcia-Molina H (2003) Designing a super-peer network. Proc International Conference on Data Engineering. Bangalore, India: 49–60.
Yi MY, Jackson JD, Park JS & Probst JC (2006) Understanding information technology acceptance by individual professionals: Toward an integrative view. Information and Management 43(3): 350–363.
Yoneki E & Bacon J (2007) Openess and interoperability in mobile middlware. In: Bellavista P & Corradi A (ed) The handbook of mobile middleware. New York, USA. Auerbach Publications: 487–518.
Yu J, Su H, Zhou G & Xu K (2007). SNet: Skip graph based semantic Web Services discovery. Proc ACM Symposium on Applied Computing. Seoul, Korea: 1393–1397.
77
Yu JY & Chong PHJ (2005) A survey of clustering schemes for mobile ad hoc networks. IEEE Communications Surveys 7(1): 32–48.
Yuan Y & Zhang JJ (2003) Towards an appropriate business model for m-commerce. International Journal of Mobile Communications 1(1/2): 35–56.
Zhao B, Huang L, Stribling J, Rhea S, Joseph A & Kubiatowicz J (2004) Tapestry: a resilient global-scale overlay for service deployment. IEEE Journal of Selected Areas in Communications 22(1): 41–53.
78
79
Original publications
I Koskela T, Kassinen O, Harjula E & Ylianttila M (In press) P2P Group Management Systems: A Conceptual Analysis. ACM Computing Surveys.
II Koskela T, Kassinen O, Ou Z & Ylianttila M (2010) Improving Community Management Performance with Two-level Hierarchical DHT Overlays. Journal of Internet Technology, Special Issue on Internet Resource Sharing and Discovery 11(2): 167–179.
III Koskela T, Harjula E, Kassinen O & Ylianttila M (2011) Robustness of a P2P Community Management System based on Two-level Hierarchical DHT Overlays. Proc IEEE Symposium on Computers and Communication. Kerkyra, Greece: 881–886.
IV Koskela T, Kostamo N, Kassinen O, Ohtonen J & Ylianttila M (2007) Towards Context-Aware Mobile Web 2.0 Service Architecture. Proc IARIA International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies. Papeete, French Polynesia: 41–48.
V Koskela T, Kassinen O, Harjula E, Pellikka J & Ylianttila M (2010) Case Study on a Community-Centric Mobile Service Environment. International Journal of Web Applications 2(3): 187–205.
VI Koskela T, Julkunen J, Kassinen O & Ylianttila M (2010) User Evaluation of a Community-centric Music Voting Service. Journal of Digital Content Technology and its Application 4(1): 69–78.
VII Koskela T, Järvinen S, Liu M & Ylianttila M (2010) User Experience in Added Value Location-based Mobile Music Service. Proc IEEE International Conference on Web Services. Miami, USA: 465–472.
Reprinted with permission from ACM (I), JIT (II), IEEE (III, IV, VII), IJWA (V)
and JDCTA (VI).
Original publications are not included in the electronic version of the dissertation.
80
A C T A U N I V E R S I T A T I S O U L U E N S I S
Book orders:Granum: Virtual book storehttp://granum.uta.fi/granum/
S E R I E S C T E C H N I C A
378. Yang, Dayou (2011) Optimisation of product change process and demand-supplychain in high tech environment
379. Kalliokoski, Juha (2011) Models of filtration curve as a part of pulp drainageanalyzers
380. Myllylä, Markus (2011) Detection algorithms and architectures for wireless spatialmultiplexing in MIMO-OFDM systems
381. Muhos, Matti (2011) Early stages of technology intensive companies
382. Laitinen, Ossi (2011) Utilisation of tube flow fractionation in fibre and particleanalysis
383. Lasanen, Kimmo (2011) Integrated analogue CMOS circuits and structures forheart rate detectors and other low-voltage, low-power applications
384. Herrala, Maila (2011) Governance of infrastructure networks : developmentavenues for the Finnish water and sewage sector
385. Kortelainen, Jukka (2011) EEG-based depth of anesthesia measurement :separating the effects of propofol and remifentanil
386. Turunen, Helka (2011) CO2-balance in the athmosphere and CO2-utilisation : anengineering approach
387. Juha, Karjalainen (2011) Broadband single carrier multi-antenna communicationswith frequency domain turbo equalization
388. Martin, David Charles (2011) Selected heat conduction problems inthermomechanical treatment of steel
389. Nissinen, Jan (2011) Integrated CMOS circuits for laser radar transceivers
390. Nissinen, Ilkka (2011) CMOS time-to-digital converter structures for theintegrated receiver of a pulsed time-of-flight laser rangefinder
391. Kassinen, Otso (2011) Efficient middleware and resource management in mobilepeer-to-peer systems
392. Avellan, Kari (2011) Limit state design for strengthening foundations of historicbuildings using pretested drilled spiral piles with special reference to St. John’sChurch in Tartu
393. Khatri, Narendar Kumar (2011) Optimisation of recombinant protein productionin Pichia pastoris : Single-chain antibody fragment model protein
394. Paavola, Marko (2011) An efficient entropy estimation approach
C396etukansi.kesken.fm Page 2 Monday, October 17, 2011 11:53 AM
ABCDEFG
UNIVERS ITY OF OULU P.O.B . 7500 F I -90014 UNIVERS ITY OF OULU F INLAND
A C T A U N I V E R S I T A T I S O U L U E N S I S
S E R I E S E D I T O R S
SCIENTIAE RERUM NATURALIUM
HUMANIORA
TECHNICA
MEDICA
SCIENTIAE RERUM SOCIALIUM
SCRIPTA ACADEMICA
OECONOMICA
EDITOR IN CHIEF
PUBLICATIONS EDITOR
Senior Assistant Jorma Arhippainen
Lecturer Santeri Palviainen
Professor Hannu Heusala
Professor Olli Vuolteenaho
Senior Researcher Eila Estola
Director Sinikka Eskelinen
Professor Jari Juga
Professor Olli Vuolteenaho
Publications Editor Kirsti Nurkkala
ISBN 978-951-42-9599-7 (Paperback)ISBN 978-951-42-9600-0 (PDF)ISSN 0355-3213 (Print)ISSN 1796-2226 (Online)
U N I V E R S I TAT I S O U L U E N S I SACTAC
TECHNICA
U N I V E R S I TAT I S O U L U E N S I SACTAC
TECHNICA
OULU 2011
C 396
Timo Koskela
COMMUNITY-CENTRIC MOBILE PEER-TO-PEER SERVICES: PERFORMANCE EVALUATION AND USER STUDIES
UNIVERSITY OF OULU,FACULTY OF TECHNOLOGY,DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING;INFOTECH OULU
C 396
ACTA
Tim
o Koskela
C396etukansi.kesken.fm Page 1 Monday, October 17, 2011 11:53 AM