Research Problems in Future-is-Here Networking Deep Medhi University - - PowerPoint PPT Presentation

research problems in future is here networking
SMART_READER_LITE
LIVE PREVIEW

Research Problems in Future-is-Here Networking Deep Medhi University - - PowerPoint PPT Presentation

Research Problems in Future-is-Here Networking Deep Medhi University of Missouri-Kansas City dmedhi@umkc.edu Switchon Workshop, October 16, 2015 International Activities US-Japan Networking Research Collaboration Started


slide-1
SLIDE 1

Research Problems in Future-is-Here Networking

Deep Medhi University of Missouri-Kansas City dmedhi@umkc.edu Switchon Workshop, October 16, 2015

slide-2
SLIDE 2

International Activities

  • US-Japan Networking Research Collaboration
  • Started in 2008 (as far as I know, this was the first one by

National Science Foundation)

  • A number of Workshops over the next few years
  • Two joint US-Japan project solicitations
  • 2009/2010: DCL (Dear Colleague Letter)
  • 2013/2014: An Open Solicitation JUNO 13-574
  • http://www.nsf.gov/pubs/2013/nsf13574/nsf13574.htm
  • Funded Project with Osaka City University
  • Brazil’s Science Without Borders Program, 2013-2015
  • Host institution: University of Campinas
  • Additional Collaborations: UFRGS, UFPR
  • France: Université Pierre et Marie Curie (UPMC), 2014-current
  • NSF SAVI travel Grant
  • iMinds – University of Ghent, University of Antwerp, Belgium
  • India: Indian Institute of Technology-Guwahati, India
slide-3
SLIDE 3

On-going Efforts

  • Software Defined Networking Scalability
  • Software Defined Networking Availability
  • Cloud Auditability: storing and mining event logs
  • Resource allocation in multi-location, multi-location data center

networks

  • Time-Dependent
  • Security-Enabled Traffic Engineering
  • IoT and Cloud and how it impacts on Core Networks
  • Hadoop in a hybrid cloud
  • Resilient Networking for Massive Failures
  • US Collaborations:
  • US: University of Kansas, Arizona State University, Duke

University

slide-4
SLIDE 4

End-to-end Resource Management for Highly Demanding Applications in Federated Wide-Area Environment

Collaborations: University of Antwerp, Belgium: Steven Latre Federal University of Rio Grande do Sol (UFRGS), Porto Alegre, Brazil: Lisandro Z. Granville University of Missouri-Kansas City (UMKC), USA

slide-5
SLIDE 5

Delivering content over a Managed IP Managed IP network

Internet ¡ Access ¡ISP ¡ Transit ¡ISP ¡ Content ¡Server ¡ Customer ¡

Managed ¡IP ¡ QoS ¡Guarantees ¡ Yes ¡ Customer ¡Base ¡ Small ¡ Service ¡Types ¡ Limited ¡

slide-6
SLIDE 6

Over Over-the-top

  • the-top (OTT) content delivery

Internet ¡ Access ¡ISP ¡ OTT ¡Provider ¡ Transit ¡ISP ¡

OTT ¡ QoS ¡ Guarantees ¡ No ¡ Customer ¡ Base ¡ Large ¡ Service ¡ Types ¡ Diverse ¡

? ¡

slide-7
SLIDE 7

Federated managed service delivery

Service ¡Provider ¡ Internet ¡ Cloud ¡site ¡ Access ¡ISP ¡ Transit ¡ISP ¡ Cloud ¡site ¡

QoS ¡guaranteed ¡ paths ¡ Dynamically ¡ deployed ¡ functions ¡

Cloud ¡site ¡

slide-8
SLIDE 8

Federated vs. Traditional

Content ¡Provider ¡ Internet ¡ Cloud-­‑Based ¡ Storage ¡Site ¡ Access ¡ISP ¡ Transit ¡ISP ¡ Content ¡Provider ¡

Managed ¡IP ¡ OTT ¡ Federated ¡ QoS ¡Guarantees ¡ Yes ¡ No ¡ Yes? ¡ Customer ¡Base ¡ Small ¡ Large ¡ Large ¡ Service ¡Types ¡ Limited ¡ Diverse ¡ Diverse ¡

slide-9
SLIDE 9

Prototype: worldwide deployment using GENI, FIRE, etc.

Kansas City, USA Porto Alegre, Brasil Antwerp, Belgium Server provider Datacenter Users

slide-10
SLIDE 10

Issues:

  • Understand the implication in really-wide area network, across

continents.

  • Application impact and rerouting: Lower level impact such as AS

graph

  • Federated Environment, where we have more control over each of

the environment

  • Understand coordination
  • Network-aware Networking (NaN)
  • Application Level characteristics for highly demanding

applications

  • Testbed setup
slide-11
SLIDE 11

Network Bandwidth Speed in a Data Center and its Impact

Collaboration: University of Campinas, Brazil: Edmundo Madeira University of Missouri-Kansas City, USA

slide-12
SLIDE 12

IEEE 802.3: Bandwidth evolution (log scale)

slide-13
SLIDE 13

Green Computing with Bw Evolution

  • Increased Network Speed = Faster VM Migrations’ Times -> can

reduce power consumption

slide-14
SLIDE 14

Impact on Makespan due to different scheduling algorithms

  • Impacts of Increased Network Speeds in Migrations, due to different

scheduling algorithms

slide-15
SLIDE 15

Impacts on Workloads with different scheduling algoritms

slide-16
SLIDE 16

Impact on VM migration

Peak # of Migrations Sharp Decline Stabilization

slide-17
SLIDE 17
  • Obtain further insight on the upper limit of the impact of

increased bandwidth consumption of energy

  • And find the factors that determine where this upper

limit is

  • Bandwidth-Aware VMs’ Scheduling Algorithms
  • Topology-Aware Schemes
  • Data Center Federation
slide-18
SLIDE 18

Acknowledgement

  • Students:
  • Federated Networking: Pedro Isolani, Shuai (Jack) Zhao,

Jeroen Famaey, Niels Soetens, Johan Bergs

  • Network Speed and its impact: Daniel Lago (UNICAMP)
  • Funding Agencies:
  • National Science Foundation Grant # 0853414, 0960801,

0916505, 1029562, 1118231, 1217736, 1302040

  • CNPq for Science Without Borders’ Program