Flowtune: Flowlet Control for Datacenter Networks
Rapid convergence to a desired allocation of network resources to endpoint traffic is a difficult problem. The reason is that congestion control decisions are distributed across the endpoints, which vary their offered load in response to changes in application demand and network feedback on a p...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
2021
|
| Online Access: | https://hdl.handle.net/1721.1/137396 |
| _version_ | 1826189931075076096 |
|---|---|
| author | Perry, Jonathan Balakrishnan, Hari Shah, Devavrat |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Perry, Jonathan Balakrishnan, Hari Shah, Devavrat |
| author_sort | Perry, Jonathan |
| collection | MIT |
| description | Rapid convergence to a desired allocation of network
resources to endpoint traffic is a difficult problem. The
reason is that congestion control decisions are distributed
across the endpoints, which vary their offered load in
response to changes in application demand and network
feedback on a packet-by-packet basis. We propose a different approach for datacenter networks, flowlet control,
in which congestion control decisions are made at the
granularity of a flowlet, not a packet. With flowlet control, allocations have to change only when flowlets arrive
or leave. We have implemented this idea in a system
called Flowtune using a centralized allocator that receives
flowlet start and end notifications from endpoints. The
allocator computes optimal rates using a new, fast method
for network utility maximization, and updates endpoint
congestion-control parameters. Experiments show that
Flowtune outperforms DCTCP, pFabric, sfqCoDel, and
XCP on tail packet delays in various settings, converging
to optimal rates within a few packets rather than over
several RTTs. Benchmarks on an EC2 deployment show
a fairer rate allocation than Linux’s Cubic. A data aggregation benchmark shows 1.61× lower p95 coflow completion time. |
| first_indexed | 2024-09-23T08:30:45Z |
| format | Article |
| id | mit-1721.1/137396 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T08:30:45Z |
| publishDate | 2021 |
| record_format | dspace |
| spelling | mit-1721.1/1373962023-03-29T19:14:04Z Flowtune: Flowlet Control for Datacenter Networks Perry, Jonathan Balakrishnan, Hari Shah, Devavrat Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Rapid convergence to a desired allocation of network resources to endpoint traffic is a difficult problem. The reason is that congestion control decisions are distributed across the endpoints, which vary their offered load in response to changes in application demand and network feedback on a packet-by-packet basis. We propose a different approach for datacenter networks, flowlet control, in which congestion control decisions are made at the granularity of a flowlet, not a packet. With flowlet control, allocations have to change only when flowlets arrive or leave. We have implemented this idea in a system called Flowtune using a centralized allocator that receives flowlet start and end notifications from endpoints. The allocator computes optimal rates using a new, fast method for network utility maximization, and updates endpoint congestion-control parameters. Experiments show that Flowtune outperforms DCTCP, pFabric, sfqCoDel, and XCP on tail packet delays in various settings, converging to optimal rates within a few packets rather than over several RTTs. Benchmarks on an EC2 deployment show a fairer rate allocation than Linux’s Cubic. A data aggregation benchmark shows 1.61× lower p95 coflow completion time. 2021-11-04T19:06:59Z 2021-11-04T19:06:59Z 2017 2019-05-02T18:02:50Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/137396 Perry, Jonathan, Balakrishnan, Hari and Shah, Devavrat. 2017. "Flowtune: Flowlet Control for Datacenter Networks." en https://www.usenix.org/system/files/conference/nsdi17/nsdi17-perry.pdf Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf MIT web domain |
| spellingShingle | Perry, Jonathan Balakrishnan, Hari Shah, Devavrat Flowtune: Flowlet Control for Datacenter Networks |
| title | Flowtune: Flowlet Control for Datacenter Networks |
| title_full | Flowtune: Flowlet Control for Datacenter Networks |
| title_fullStr | Flowtune: Flowlet Control for Datacenter Networks |
| title_full_unstemmed | Flowtune: Flowlet Control for Datacenter Networks |
| title_short | Flowtune: Flowlet Control for Datacenter Networks |
| title_sort | flowtune flowlet control for datacenter networks |
| url | https://hdl.handle.net/1721.1/137396 |
| work_keys_str_mv | AT perryjonathan flowtuneflowletcontrolfordatacenternetworks AT balakrishnanhari flowtuneflowletcontrolfordatacenternetworks AT shahdevavrat flowtuneflowletcontrolfordatacenternetworks |