TOP
Search the Dagstuhl Website
Looking for information on the websites of the individual seminars? - Then please:
Not found what you are looking for? - Some of our services have separate websites, each with its own search option. Please check the following list:
Schloss Dagstuhl - LZI - Logo
Schloss Dagstuhl Services
Seminars
Within this website:
External resources:
  • DOOR (for registering your stay at Dagstuhl)
  • DOSA (for proposing future Dagstuhl Seminars or Dagstuhl Perspectives Workshops)
Publishing
Within this website:
External resources:
dblp
Within this website:
External resources:
  • the dblp Computer Science Bibliography


Dagstuhl Seminar 16281

Network Latency Control in Data Centres

( Jul 10 – Jul 13, 2016 )

(Click in the middle of the image to enlarge)

Permalink
Please use the following short url to reference this page: https://www.dagstuhl.de/16281

Organizers

Contact



Motivation

Data centres are at the heart of the modern Internet. They host web services, social networking, cloud computing and are increasingly used by operators to host virtual network functions. All these services have one thing in common: they require extremely low latency communication in the data centre. Consequently we have seen the birth of a new field in networking research – data centre latency control.

Unlike the earlier generation of high-performance computing clusters, data centres have tended to use commodity off-the-shelf servers and switches, and run standard operating systems. However, traditional networking equipment and TCP-IP stacks were designed for wide-area networks, where the goal is to maximize throughput, and the control loop between end systems is measured in 10s of milliseconds. By contrast, data centres operate on timescales that are several orders of magnitude lower. And while throughput is important, the plentiful bandwidth of data centre networks makes throughput a secondary concern to latency.

This seminar will explore existing and future techniques for controlling data centre latency across the entire software and hardware stack, including in-network solutions, end-host solutions, and others. The aims of the seminar are to foster closer collaboration between academic researchers, industry, and operators.

Each participant is encouraged to present their thoughts or work on the subject in the form of short "lightning" talks. The majority of the seminar time will be used for discussions, breakout sessions and other ways that we hope will lead to new collaborative ideas and approaches in this space, or interesting refinements to existing ideas.

We are inviting people with a number of different backgrounds, from academia to open source and commercial developers, to data centre operators, application, transport, security and switching/routing exerts. We intend to document the results of the seminar in a joint publication, to which every participant is invited to contribute.

If you have not been at Dagstuhl before: Dagstuhl is a unique place in the middle of nowhere, about 2 hours away from Frankfurt airport. The research centre is located in an 18th century castle with an extension building from 1990. Dagstuhl was set up as a seminar centre for current research topics in computer science that provides the infrastructure needed to host successful seminars, offering rooms, meals, a fast Internet connection, and an exceptional library. More information is available at http://www.dagstuhl.de/en/about-dagstuhl/.


Summary

Data centres are at the heart of the modern Internet. They host web services, social networking, cloud computing and are increasingly used by operators to host virtual network functions. All these services have one thing in common: they require extremely low latency communication in the data centre. Consequently we have seen the birth of a new field in networking research - data centre latency control.

Unlike the earlier generation of high-performance computing clusters, data centres have tended to use commodity off-the-shelf servers and switches, and run standard operating systems. However, traditional networking equipment and TCP-IP stacks were designed for wide-area networks, where the goal is to maximize throughput, and the control loop between end systems is measured in 10s of milliseconds. By contrast, data centres operate on timescales that are several orders of magnitude lower. And while throughput is important, the plentiful bandwidth of data centre networks makes throughput a secondary concern to latency.

This seminar explored existing and future techniques for controlling data centre latency across the entire software and hardware stack, including in-network solutions, end-host solutions, and others. The aims of the seminar are to foster closer collaboration between academic researchers, industry, and operators. 38 researchers attended the multidisciplinary seminar. Over the course of the 3-day seminar, seven presentations were given on various aspects of data center networking. Taking the presentations as input, the workshop then broke into six working groups to discuss research aspects of latency control. The seminar was concluded by voting and discussing on possible conclusions from our discussions. Each conclusion was discussed briefly, then voted on. The outcome of the breakout session as well as the concluding statements are summarized in this report.

Copyright Oliver Hohlfeld and Mohammad Alizadeh Attar and Jon Crowcroft and Lars Eggert and Klaus Wehrle

Participants
  • Alexandru Agache (University Politehnica of Bucharest, RO) [dblp]
  • Mohammad Alizadeh Attar (MIT - Cambridge, US) [dblp]
  • Jari Arkko (Ericsson - Jorvas, FI) [dblp]
  • Marinho Barcellos (Federal University of Rio Grande do Sul, BR) [dblp]
  • Olivier Bonaventure (University of Louvain, BE) [dblp]
  • Bob Briscoe (Simula Research Laboratory - Lysaker, NO) [dblp]
  • Edouard Bugnion (EPFL Lausanne, CH) [dblp]
  • Marco Canini (University of Louvain, BE) [dblp]
  • Julian Chesterfield (University of Cambridge, GB) [dblp]
  • Mosharaf Chowdhury (University of Michigan - Ann Arbor, US) [dblp]
  • Jon Crowcroft (University of Cambridge, GB) [dblp]
  • Koen De Schepper (NOKIA Bell Labs - Antwerp, BE) [dblp]
  • Aaron Ding (TU München, DE) [dblp]
  • Nandita Dukkipati (Google Inc. - Mountain View, US) [dblp]
  • Lars Eggert (NetApp Deutschland GmbH - Kirchheim, DE) [dblp]
  • Patrick Thomas Eugster (TU Darmstadt, DE) [dblp]
  • Matthew P. Grosvenor (University of Cambridge, GB) [dblp]
  • Dongsu Han (KAIST - Daejeon, KR) [dblp]
  • Oliver Hohlfeld (RWTH Aachen, DE) [dblp]
  • Michio Honda (NetApp Deutschland GmbH - Kirchheim, DE)
  • Patrick Jahnke (TU Darmstadt, DE) [dblp]
  • Lavanya Jose (Stanford University, US) [dblp]
  • Wolfgang Kellerer (TU München, DE) [dblp]
  • Kirill Kogan (IMDEA Networks - Madrid, ES) [dblp]
  • Mirja Kühlewind (ETH Zürich, CH) [dblp]
  • Mike Marty (Google - Madison, US)
  • Andrew W. Moore (University of Cambridge, GB) [dblp]
  • John Ousterhout (Stanford University, US) [dblp]
  • Jitendra Padhye (Microsoft Corporation - Redmond, US) [dblp]
  • Rong Pan (CISCO Systems - San Jose, US) [dblp]
  • Max Plauth (Hasso-Plattner-Institut - Potsdam, DE) [dblp]
  • Ihsan Ayyub Qazi (LUMS - Lahore, PK) [dblp]
  • Costin Raiciu (University Politehnica of Bucharest, RO) [dblp]
  • Michael Scharf (NOKIA - Stuttgart, DE) [dblp]
  • Hakim Weatherspoon (Cornell University - Ithaca, US) [dblp]
  • Klaus Wehrle (RWTH Aachen, DE) [dblp]
  • Michael Welzl (University of Oslo, NO) [dblp]
  • Noa Zilberman (University of Cambridge, GB) [dblp]

Classification
  • networks
  • operating systems
  • optimization / scheduling

Keywords
  • Network architecture
  • data centres
  • latency
  • resource control
  • scheduling
  • end-to-end transport protocols.