Network Softwarization: Trends, Challenges, and Research Efforts

Encontro Telecom @ Unicamp

Prof. Christian Esteve Rothenberg

Department of Computer Engineering and Industrial Automation (DCA)

Faculty of Electrical and Computer Engineering (FEEC)

University of Campinas (UNICAMP)

Department of Computer Engineering and Industrial Automation (DCA)

Faculty of Electrical and Computer Engineering (FEEC)

University of Campinas (UNICAMP)

• Very intensive in hardware

• Software not at the core

• Very intensive in software

• Hardware is a necessary base

x

HARDWARE SOFTWARE+

-

+

-

Traditional telcosTraditional telcos Internet playersInternet players

Adapt to survive: Telco evolution focus shifting from hardware to software

Source: Adapted from D. Lopez Telefonica I+D, NFV

Enter the Software-Defined Era

Google, FacebookAT&T, Telefonica, Telebras

NFV >>> Accelerating Transformation

Source: Adapted from D. Lopez Telefonica I+D, NFV

Trends• High performance industry

standard servers shipped in very high volume

• Convergence of computing, storage and networks

• New virtualization technologiesthat abstract underlying hardware yielding elasticity, scalability and automation

• Software-defined networking

• Cloud services

• Mobility, explosion of devices and traffic

Challenges• Huge capital investment to deal with

current trends

• Network operators face an increasing disparity between costs and revenues

• Complexity: large and increasing variety of proprietary hardware appliances in operator’s network

• Reduced hardware lifecycles

• Lack of flexibility and agility: cannot move network resources where & when needed

• Launching new services is difficult and takes too long. Often requires yet another proprietary box which needs to be integrated

Source: Adapted from D. Lopez Telefonica I+D, NFV

A means to make the network more flexible and simple by minimising dependence on HW constraints

The NFV Concept

Source: Adapted from D. Lopez Telefonica I+D, NFV

BRAS

FirewallDPI

CDN

Tester/QoEmonitor

WANAcceleration

MessageRouter

Radio AccessNetwork Nodes

CarrierGrade NAT

Session BorderController

Network Virtualisation Approach

PE RouterSGSN/GGSN

Independent Software Vendors

Standard High VolumeEthernet Switches

Standard High Volume Storage

Standard High Volume Servers

Orchestrated,automatic &remote install.

Classical Network Appliance Approach

SDN & NFVIT & Networking Growing Together

Source: NEC

Network equipment as Black boxes

Open interfaces (OpenFlow) for instructing the boxes what to do

SDN

Boxes with autonomousbehaviour Decisions are taken out of the box

FEATURE FEATURE

OPERATING SYSTEM

SPECIALIZED PACKET FORWARDING HARDWAREFEATURE FEATURE

OPERATING SYSTEM

SPECIALIZED PACKET FORWARDING HARDWARE

FEATURE FEATURE

OPERATING SYSTEM

SPECIALIZED PACKET FORWARDING HARDWAREFEATURE FEATURE

OPERATING SYSTEM

SPECIALIZED PACKET FORWARDING HARDWARE

SDN

Adapting OSS to manage black boxesSimpler OSS to manage the SDN controller

SDNFEATURE FEATURE

OPERATING SYSTEM

SPECIALIZED PACKET FORWARDING HARDWAREFEATURE FEATURE

OPERATING SYSTEM

SPECIALIZED PACKET FORWARDING HARDWARE

FEATURE FEATURE

OPERATING SYSTEM

SPECIALIZED PACKET FORWARDING HARDWAREFEATURE FEATURE

OPERATING SYSTEM

SPECIALIZED PACKET FORWARDING HARDWARE

Software Defined Networking

Source: Adapted from D. Lopez Telefonica I+D, NFV

NFV vs SDN

• NFV: re-definition of network equipment architecture

• NFV was born to meet Service Provider (SP) needs:

• Lower CAPEX by reducing/eliminating proprietary hardware

• Consolidate multiple network functions onto industry standard platforms

• SDN: re-definition of network architecture

• SDN comes from the IT world:

• Separate the data and control layers, while centralizing the control

• Deliver the ability to program network behavior using well-defined interfaces

NFV vs SDN

SDN ››› flexible forwarding & steering of trafficin a physical or virtual network environment

NFV ››› flexible placement of virtualizednetwork functions across the network & cloud

››› SDN & NFV are complementary tools for achieving full network programmability

Flexibility with SDN & NFV

Source: Ahmad Rostami, Ericsson Research (Kista): http://www.itc26.org/fileadmin/ITC26_files/ITC26-Tutorial-Rostami.pdf

Source: Bob Briscoe, BT

So, why we need/want NFV(/SDN)?1. Virtualization: Use network resource without worrying about where it is physically located, how much it is, how it is organized, etc.

2. Orchestration: Manage thousands of devices

3. Programmable: Should be able to change behavior on the fly.

4. Dynamic Scaling: Should be able to change size, quantity, as a F(load)

5. Automation: Let machines / software do humans’ work

6. Visibility: Monitor resources, connectivity

7. Performance: Optimize network device utilization

8. Multi-tenancy: Slice the network for different customers (as-a-Service)

9. Service Integration: Let network management play nice with OSS/BSS

10. Openness: Full choice of modular plug-ins

Note: These are exactly the same reasons why we need/want SDN.Source: Adapted from Raj Jain

The Frontier of Networking

Existing

• CLIs

• Closed Source

• Vendor Lead

• Classic Network Appliances

New

• APIs

• Open Source

• Customer Lead

• Network Function Virtualization (NFV)

Adapted from: Kyle Mestery, Next Generation Network Developer Skills

• Assistant Professor (tenure track) at FEEC/UNICAMP (since 2013)

• Leading the INTRIG lab at DCA/FEEC/UNICAMPINTRIG: Information & Networking Technologies Research & Innovation Group

• Currently, supervising 5 PhD, 5 MSc candidates, and 4 undergrad students

• PhD in Electrical and Computer Engineering (FEEC/UNICAMP, 2010), MSc in Electrical Eng and Information Technology (Darmstadt University, 2006), Telecommunication Eng (Universidad Politécnica de Madrid, 2004)

• Visiting researcher at Ericsson Research Nomadic Lab, Jorvas, Finland, 2008, participated in EU Publish/Subscribe Internet Routing Paradigm (PSIRP).

• Research Scientist at CPqD R&D Center in Telecommunication (2010-2013)

• Technical Lead of SDN activities in the Converged Networking Division

• ONF Research Associate (since Apr/2013)

About: Christian Esteve Rothenberg

Research Interests and Main Goals & Results

Software

Defined

Networking

Information

Centric

Networking

Network

Functions

Virtualization

• RouteFlow(hybrid IP-SDN)

• softswitch13

• libfluid(ONF Driver)

• SDN-2-SDN Peering

• MD2-NFV(Multi-Domain Distributed)

• VNF-a-a-S

• Mini-CCNx -> Mini-NDNx

• LIPSIN(in-packet Bloom filters)

• 10G VNF

• Multi-Core Architectures

• Mininet-WiFi :: Wireless / Mobility

Technical lead of successful open source projects:

• libfluid, winner of the ONF Driver Competition (Mar/2014)

• http://opennetworkingfoundation.github.io/libfluid/

• softswitch13, first OpenFlow 1.2 and 1.3 soft switch, controller, and testing framework [funded and in technical collaboration with Ericsson] (2011 - 2013)

• https://github.com/CPqD/ofsoftswitch13

• Mini-CCNx, fast prototyping and experimentation of CCN networks (2013 - )

• https://github.com/carlosmscabral/mn-ccnx

• RouteFlow, first IP routing architecture for SDN (2010 - )

• https://github.com/routeflow/

More info: https://github.com/intrig-unicamp/ https://github.com/chesteve/16

Research Projects and Open Source Results

More info: Mininet-WiFi: Emulating Software-Defined Wireless Networkshttps://github.com/intrig-unicamp/mininet-wifi

17

Mininet-WiFi

More info: DemosVideo 01: https://www.youtube.com/watch?v=_PtSmhf7Z8sVideo 02: https://www.youtube.com/watch?v=H46EPuJDJhcVideo 03: https://www.youtube.com/watch?v=WH6bSOKC7Lk

Mininet-WiFi :: Use Cases

19

Mininet-WiFi :: Research TargetTowards a high-fidelity wireless emulator

• Duration: 24 months (Started in May/2014) [Funded by Ericsson]

20

Ongoing research project:Advanced Peering with a Software-Defined Knowledge Plane

MD2-NFV

• Resource trading (computation, storage, network) between different domains to optimize VNF placement across providers (offering NFVI-as-a-Service)

• More info: "MD2-NFV: The Case for Multi-Domain Distributed Network Functions Virtualization". In NetSys SDNFlex 2015.

Flash NFV-Benchmarking Service

• A fast, distributed benchmarking service to assess candidate locations in terms of computation and network (BW, latency, etc.) from multiple vantage points to support the decision of best location for target VNFs (“try before deploy”).

• More info: "VBaaS: VNF Benchmark-as-a-Service". In fourth edition of the European Workshop on Software Defined Networks (EWSDN 2015).

KP-Query

• Network State & Management primitives for graph databases embodying multi-layer virtualized infrastructures based on semantic graph annotations

• More info: "Towards Semantic Network Models via Graph Databases for SDN Applications". In fourth edition of the European Workshop on Software Defined Networks (EWSDN 2015).

21

Related use cases under investigation:Advanced Peering with a Software-Defined Knowledge Plane

22

Mapping the Big Picture to Partial Results & ResearchAdvanced Peering with a Software-Defined Knowledge Plane

• Semantic Models / APIs based onNML (Network Markup Language)

• Knowledge generation based onMachine Learning & Big Data

• ALTO as East/West “Peering” APIs between SDN domains• Exchange Topology and Cost Maps• Optimize Placement & Routing towards Services/Content

• Anatomy of current IXP Ecosystem

• Extracting requirements for SDN peering andrealistic topologies for inter-domain SDN

Location of Each Brazilian IXP (figure extracted from http://www.ptt.br)

023

PTTMetro in Brazil (PTT.br)

Figure 1b. Example Graph of PTT-VIX (Vitória, ES)

Data Sources

• PTT.br Official Data at www.ptt.br

• PeeringDB (we found it was unreliable)

• Telnet Access to IXP's Looking Glasses (BGP Table, Paths Summary, Communities List

Graph analysis on BGP adjacency matrix of all IXPs using NetworkX and Neo4j)

Methodology: AS-level Graphs

25Results (see details in SBRC´15 & on-demand)

• Profile and Classification of IXP's Members

• Who is currently where? Who should be there?

• AS Vertices's Degree / Depth / Diameter

• How many? How far?

• Density of Peering

• How many unexpoited peering relationships? –> Peering Recommender system–> SDN-enabled peering opportunities

• AS-Prepend for TE

• How much indirect TE through BGP hacks?–> SDN-enabled QoS between peers

• k-Clique Communities & Machine Learning (ML)

• Organization and Classification of Customer/Client/Peers

• –> Peering Recommender system?

Thanks! Obrigado! (More) Questions?

chesteve@dca.fee.unicamp.br

http://www.dca.fee.unicamp.br/~chesteve/

http://www.intrig.dca.fee.unicamp.br

26

• Duration: 24 months (Started in May/2014) [Funded by Ericsson]

27

Ongoing research project:Advanced Peering with a Software-Defined Knowledge Plane

• Investigate SDN-SDN communication options that allow SDN networks in different administrative domains to achieve advanced peering agreements beyond pure packet routing, for instance by integrating best placement of content and applications.

• Re-examine the concept of Knowledge Plane with the visibility and SDN abstractions (topology map, flow tuples, policy specification) and direct control capabilities (i.e., rich matching and instructions) of OpenFlow/SDN approaches.

• Design Knowledge Plane mechanisms for querying (SDN) network properties of participants at IXPs. Define APIs that allow operators to differentiate and close the gap between applications and networks, with focus on optimizing content/application placement.

• Develop novel SDN control loops (information gathering + knowledge generation + network actuation) leveraging SDN visibility (among other data sources) for data collection and network control via direct OpenFlow/SDN programmability or indirect via existing protocols (non-OF SDN).

• Design suitable data models and available graph-oriented DBs (and/or big data stores for map-reduce like operations) to embody the "Network Information Base” using annotated and semantically meaningful graphs.

• Prototypes of inter-SDN architecture and its building blocks for evaluation and proof of concept purposes. Software prototyping based on OpenDaylight, NoSQL graph-oriented databases (e.g. Neo4j, Titan), and Hadoop-like infrastructures.

Research GoalsAdvanced Peering with a Software-Defined Knowledge Plane

• Christian Esteve Rothenberg is an Assistant Professor in the Faculty of Electrical and Computer Engineering at University of Campinas (UNICAMP), where he received his Ph.D. in Computer Engineering in 2010.

• From 2010 to 2013, he worked as Senior Research Scientist in the areas of IP systems and networking at CPqD Research and Development Center in Telecommunications (Campinas, Brazil), where he was technical lead of R&D acitivities in the field of OpenFlow and SDN such as the RouteFlow project, the OpenFlow 1.3 Ericsson/CPqDsoftswitch, or the ONF Driver competition.

• Christian holds the Telecommunication Engineering degree from Universidad Politécnicade Madrid (ETSIT - UPM), Spain, and the M.Sc. (Dipl. Ing.) degree in Electrical Engineering and Information Technology from the Darmstadt University of Technology (TUD), Germany, 2006.

• Christian holds two international patents and has over 70 publications including scientific journals and top-tier networking conferences such as SIGCOMM and INFOCOM. Since April 2013, Christian is an Open Networking Foundation (ONF) Research Associate.

29

Short-bio

• Software-Defined Networking: A Comprehensive Survey. Diego Kreutz, Fernando M. V. Ramos, Paulo Verissimo, Christian Esteve Rothenberg, Siamak Azodolmolky, Steve Uhlig. In Proceedings of the IEEE, Jan., 2015.

• http://arxiv.org/abs/1406.0440 https://github.com/SDN-Survey/latex/wiki

• When Open Source Meets Network Control Planes. Rothenberg, C.E. Chua, R. ; Bailey, J. ; Winter, M. ; Correa, C.N.A. ; de Lucena, S.C. ; Salvador, M.R. ; Nadeau, T.D.. In IEEE Computer, vol.47, no.11, pp.46,54, Nov. 2014

• Cardigan: SDN Distributed Routing Fabric Going Live at an Internet Exchange.Jonathan P. Stringer, Carlos Corrêa, Josh Bailey, Dean Pemberton, Qiang Fu, Christopher Lorier, Richard Nelson, Christian Esteve Rothenberg. In IEEE ISCC, June 2014

• Cardigan: Deploying a Distributed Routing Fabric." In ACM SIGCOMM 2013 - HotSDN'13 (Poster)

• Revisiting Routing Control Platforms with the Eyes and Muscles of Software-Defined NetworkingChristian E. Rothenberg, Marcelo R. Nascimento, Marcos R. Salvador, Carlos Corrêa, Sidney Lucena, and Robert Raszuk. In ACM SiGCOMM HotSDN, Aug 2012

30

(Selected) Publications related to SDN

• Marcos Siqueira, Fabian Hooft, Juliano Oliveira, Edmundo Madeira, Christian Esteve Rothenberg, Providing Optical Network as a Service with Policy-based Transport SDN. In Journal of Network and Systems Management, June 2014.

• P. Jokela, H. Mahkonen, C. Esteve Rothenberg, and J. Ott. (Deployable) Reduction of Multicast State with In-packet Bloom Filters. In IFIP NETWORKING, 2013.

• S. Tarkoma, C. Esteve Rothenberg and E. Lagerspetz. Theory and Practice of Bloom Filters for Distributed Systems. In IEEE Communications Surveys and Tutorials. Vol. 14, Number 1, 2012

• M. Särelä, C. Esteve Rothenberg, T. Aura, A. Zahemszky, P. Nikander and J. Ott. Forwarding Anomalies in Bloom Filter Based Multicast, In IEEE INFOCOM 2011.

• P. Jokela, A. Zahemszky, C. Esteve, S. Arianfar, and P. Nikander. LIPSIN: Line speed Publish/Subscribe Inter-Networking. In ACM SIGCOMM 2009.

31

(Selected) Publications

More info: "MD2-NFV: The Case for Multi-Domain Distributed Network Functions Virtualization". To appear in NetSys SDNFlex 2015, Cottbus, Germany, March, 2015. 32

Peering Use Cases in Multi-Domain Distributed NFV

33

Results

34

Results

35

Results

36

Results

37

Results