Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario
-
Upload
daniela-mazza -
Category
Engineering
-
view
340 -
download
4
description
Transcript of Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario
![Page 1: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/1.jpg)
Challenges on Wireless HetNet for Mobile Cloud Computing
in a Smart City scenario Bologna, November 7th 2014
Daniela Mazza, PhD Student - 28th Cycle
Department of Electronics Engineering, Telecommunications and Information Technology University of Bologna, Italy Supervisor: Prof. Giovanni Emanuele CorazzaCo-advisor: Prof. Daniele Tarchi
![Page 2: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/2.jpg)
Outline◆ Urbanization and ICT trends. The Smart City concept ◆ Urban Mobile Cloud Computing
◆ HetNets: Macro and small cells ◆ Cloud Topologies
◆ Offloading in UMCC: Throughput, Energy and Time spent for computation
◆ Cost Function ◆ Numerical results
![Page 3: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/3.jpg)
Urbanization: where are we?
Source: United Nations World Urbanization Prospects 2014 Revision
![Page 4: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/4.jpg)
Urbanization: where are we?
2014: 28 mega-cities (>10M inhabitants) 54% of population resides in urban area
Source: United Nations World Urbanization Prospects 2014 Revision
![Page 5: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/5.jpg)
Urbanization: where are we going?
2030: 41 mega-cities (>10M inhabitants) 60% of population resides in urban area
Source: United Nations World Urbanization Prospects 2014 Revision
![Page 6: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/6.jpg)
Urbanization: Where are we going?
![Page 7: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/7.jpg)
Societal Challenges
Energy supply, waste management, natural disasters, energy consumption, traffic, pollution, …….
![Page 8: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/8.jpg)
Global Mobile vs Desktop Internet User Projection (Morgan Stanley Research)
Connections: Where are we?
![Page 9: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/9.jpg)
Connections: where are we going?
Cisco VNI Forecast• 2018: almost 4 billion Internet users, 52% of the world’s projected
population. • the average fixed broadband speed will grow from 16 to 42 Mbps from
2013 to 2018
287M → 317M
235M → 371M
323M → 346M 224M → 431M
213M → 431M
1.2B → 2.1B
![Page 10: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/10.jpg)
Smart CityA city that promotes the use of ICT to make better use of infrastructure, reduces the use of environmental capital and supports smart growth, to achieve a better urban way of life..
• Environment-friendly design buildings • Regional Emergency Medical Service • Smart Buildings • MegaSolar • Biomass Fuels • Electric Vehicle Car Sharing • Smart House • Electric Bus • Multi-energy Station • Off-shore wind farm • Solar panel • Intelligent transportation System (ITS) • Next Generation vehicle center • Battery Storage System • WindFarm
this image: 197 results on Google
“Smart City”: 250.000.000 results on
![Page 11: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/11.jpg)
Smart City and data exchange
System of systems (main functional areas interconnected)
Data exchanged (Users devices as data input / output )
Wireless Communication – data are exchanged between the citizens' devices and the Smart City system both uploading and downloading
![Page 12: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/12.jpg)
Smart City and data exchange
• Sensors: acquisition of data regarding the users and the environment
• Nodes: organization of a distributed mobile cloud, VCN (Vehicular Cloud Network)
• Outputs: providing results for users and for machines (M2M)
![Page 13: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/13.jpg)
Urban Mobile Cloud Computing Framework
Urban area with a pervasive wireless coverage, where several mobile devices are interacting with:
• a traditional centralized cloud service
• roadside units (cloudlets)
• a distributed mobile cloud consisting of many SMD
Access nodes of the HetNet (macro and microcells) connecting SMD to the Centralized Cloud
![Page 14: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/14.jpg)
Cloud TopologiesCentralized Cloud (remote infrastructure) • big storage capacity • high computing power • elasticity of resource provisioning • drawbacks: latency, congestion
Distributed Mobile Cloud (neighboring SMD sharing resouces) • small storage capacity (each SMD) • small computing power (each SMD) • useful when neighbors need the same
resources
Cloudlets (proximity infrastructures) • medium storage capacity • medium computing power • address latency drawbacks • drawbacks: limited area
![Page 15: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/15.jpg)
HetNet: Macro and small cellsMacrocells (3G, LTE):
• coverage > 500 m • total coverage of the area • minimal handover frequency • channel fading and traffic congestion
Small cells Picocells (malls, airports, stadium):
• coverage > 200m • High number of connected devices
Femtocells (home or small business): • coverage < 200m • Only for selected devices
WiFi access (home or small business): • Coverage < 100 m • Only for selected devices
![Page 16: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/16.jpg)
Application Requirements
APPLICATIONS latency energy throughput
computing
exchanged data
storage users
Mobility restrictive variable restrictive high high variable high
Healthcare restrictive non-restrictive
non-restrictive high high high low
Disaster Recovery restrictive restrictive non-restrictive high high high variable
Energy non-restrictive
non-restrictive
non-restrictive high high high high
Waste Management non-restrictive restrictive non-
restrictive low low low low
Tourism non-restrictive restrictive non-
restrictive high high high variable
![Page 17: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/17.jpg)
System Interactions
The utility function acts for distributing and performing the application in different parts of the Urban MCC
Devices and clouds
Processing speed
Storage Capacity
Communication equipments
Channel capacity
Priority QoS management
Communication Interfaces
QoS Requirements
Latency
Energy consumption
Throughput
Computing
Exchanged data
Storage
Users
Smart City Applications
Mobility
Healthcare
Disaster recovery
Energy
Waste Management
Tourism
Utility or Cost
Function
Partition of the
application and node and cloud
association
![Page 18: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/18.jpg)
System Interactions
![Page 19: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/19.jpg)
Offloading Distribution among the different topologies of clouds
![Page 20: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/20.jpg)
Througput
BW bandwidth n no. of the devices connected to the node
SNR Signal to Noise Ratio
d (distance from the device to the node)
d
n
![Page 21: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/21.jpg)
Energy and time for computationUser's point of view • The mobile device consumes energy
to transfer data to the cloud • The mobile device consumes (little)
energy waiting for the computation while the task is performed in the cloud
• The mobile device consumes energy to transfer results from the cloud
● The mobile device consumes energy for the computation of the task
● The time is related to the trasfer of data from the mobile device and transfer of results from the cloud
● The computation is faster due to the high computing capacity of the cloud servers
● The time is related to the poor computing capacity of the mobile device
![Page 22: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/22.jpg)
Local computation
C number of instructions of the task
Smd calculation speed Pl power for local computing
Energy for local computation: Time for local computation:
![Page 23: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/23.jpg)
Total data offloading
Cloud server computation
D exchanged data Ptr power for sending and receiving data Str transmission speed
C instructions (no.)
Pid power while being idle
Scs cloud server’s calculation speed
Energy for total offloading computing: Time for total offloading computing:
![Page 24: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/24.jpg)
Partial offloadingLocal
computationOffloading data Cloud server
computation
C instructions (no.)D exchanged data (bit)
C instructions (no.)
weight coefficients - percentage of the computational task and of the exchanged data for offloading
![Page 25: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/25.jpg)
Cost Function
Network centric approach bounded discretionary chosen (= 0.5)
![Page 26: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/26.jpg)
Numerical results
LTE eNodeB – channel capacity 100 mHz
WiFi acces points – channel capacity 22 mHz
Pid = 0.3 W Power while being idle
Smd = 400 MHz Computation Speed
Pl = 0.9 W Power for local computing
Ptr = 1.3 W Power for sending and receiving data
![Page 27: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/27.jpg)
Numerical results
Application 1: Real time traffic analysis
Application 2: mobile video and audio communication
Application 3: mobile social networking
When the network is overloaded,, with both a large amount of computation to execute and data to exchange, tasks are better performed for a specific value of gamma
![Page 28: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/28.jpg)
Application 3 – Cost function's results
Energy and time consumption for the application with high computation and high amount of data to be transferred
![Page 29: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/29.jpg)
A User-Satisfaction Based Utility Function
U1(x) =1
1+ e−α (x−β )U2 (x) = 1−
11+ e−α (x−β )
f2 (Epart _od ,ijk ) = 1−1
1+ e−α2 (Epart _od ,ijk−Eo ,k )f1(Str ,ij ) =
11+ e−α1(Str ,ij−Stro ,k )
f3(Tpart _od ,ijk ) = 1−1
1+ e−α3 (Tpart _od ,ijk−To ,k )
Uij = c1 ⋅ f1(Str ,ij )+ c2 ⋅ f2 (Epart _od ,ijk )+ c3 ⋅ f3(Tpart _od ,ijk )
![Page 30: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/30.jpg)
Reference Values
![Page 31: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/31.jpg)
Numerical Results
Performance results in terms of average energy consumption with a variable number of SMDs
![Page 32: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/32.jpg)
Numerical Results
Performance results in terms of average computation time with a variable number of SMDs
![Page 33: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/33.jpg)
Numerical Results
Performance results in terms of average throughput time with a variable number of SMDs
![Page 34: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/34.jpg)
Complexity• M available HetNet nodes Nod[i] for offloading towards the centralized
cloud,
• N cloudlets Ccl[j]
• K SMDs MD[k], to share the computation in the distributed cloud
• total of 1 + M + N + K entities, including the local node RSMD
Aim: to distribute, by means of all these entities, different percentages αi of operations O, βi of data D, and γi of memory S, to all the available nodes, cloudlets and SMDs.
![Page 35: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/35.jpg)
A real application: realtime navigation
Cloudlets only
Cloudlets and near vehicle
![Page 36: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/36.jpg)
Numerical Results
![Page 37: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/37.jpg)
Numerical Results
![Page 38: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/38.jpg)
Numerical Results
![Page 39: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/39.jpg)
Numerical Results
![Page 40: Challenges on wireless Heterogeneous Networks for Mobile Cloud Computing in a Smart City scenario](https://reader031.fdocuments.in/reader031/viewer/2022020123/559c3b7b1a28abdb7f8b45fc/html5/thumbnails/40.jpg)
PapersD. Mazza, D. Tarchi, and G. E. Corazza, “A partial offloading technique for wireless mobile cloud computing in smart cities,” in Proc. of 2014 European Conference on Networks and Communications (EuCNC), Bologna, Italy, Jun. 2014.
D. Mazza, D. Tarchi, and G. E. Corazza, “A user-satisfaction based offloading technique for smart city applications,” in Proc. of IEEE Globecom 2014, Austin, TX, USA, Dec 2014, accepted for publication.
D. Mazza, D. Tarchi, and G. E. Corazza,, “Urban mobile cloud computing: a framework at the service of smart cities,” IEEE Commun. Mag., submitted.
D. Mazza, D. Tarchi, and G. E. Corazza, “Improving Execution of Smart City Applications Through Heterogeneous Networks and Clouds,” IEEE ICC International Conference on Communication 2015, London, UK, submitted.