• Corpus ID: 15319361

Queues Don't Matter When You Can JUMP Them!

@inproceedings{Grosvenor2015QueuesDM,
  title={Queues Don't Matter When You Can JUMP Them!},
  author={Matthew P. Grosvenor and Malte Schwarzkopf and Ionel Gog and Robert N. M. Watson and Andrew W. Moore and Steven Hand and Jon Crowcroft},
  booktitle={NSDI},
  year={2015}
}
QJUMP is a simple and immediately deployable approach to controlling network interference in datacenter networks. [] Key MethodWe have implemented QJUMP as a Linux Traffic Control module. We show that QJUMP achieves bounded latency and reduces in-network interference by up to 300×, outperforming Ethernet Flow Control (802.3x), ECN (WRED) and DCTCP. We also show that QJUMP improves average flow completion times, performing close to or better than DCTCP and pFabric.

Figures from this paper

ProCAM: A Proactive Coordinating Mechanism for Low-Congestion Datacenter Networks
  • Xin Xu, Wei Zhou, Jianguo Yao
  • Computer Science
    2016 IEEE 18th International Conference on High Performance Computing and Communications; IEEE 14th International Conference on Smart City; IEEE 2nd International Conference on Data Science and Systems (HPCC/SmartCity/DSS)
  • 2016
TLDR
The extensive simulations show that the proposed ProCAM can achieve high throughput with nearly zero packet loss and low latency when the network is heavily loaded, and queue length in buffers can be maintained at a low level and the performance of latency-sensitive flows can be guaranteed.
Latency-based Congestion Detection and Control for Datacenters Major Advisor :
TLDR
This dissertation explores latency as accurate congestion feedback and develops a new congestion control algorithm that is specifically designed to achieve low queueing delay in datacenter networks.
Protocol-Independent Service Queue Isolation for Multi-Queue Data Centers
  • Gyuyeong Kim, Wonjun Lee
  • Computer Science, Mathematics
    2020 IEEE 40th International Conference on Distributed Computing Systems (ICDCS)
  • 2020
TLDR
DynaQ is designed, a protocol-independent multi-queue management solution that enables service queue isolation with generic transport protocols and dynamically adjusts the packet dropping threshold of service queues.
Exploring Token-Oriented In-Network Prioritization in Datacenter Networks
TLDR
This article proposes a readily-deployable remedy to achieve in-network prioritization by pushing both switch and end-host hardware capacity to an extreme end and implements a running TOP system with Linux hosts and commodity switches, and investigates the applicability of TOP.
TFC: token flow control in data center networks
TLDR
An clean-slate explicit transport control mechanism, called Token Flow Control (TFC), is proposed for data center networks to achieve high link utilization, ultra-low latency, fast convergence, and rare packets dropping.
One more queue is enough: Minimizing flow completion time with explicit priority notification
TLDR
Explicit Priority Notification (EPN) is the first system that provides millions of priorities for flow scheduling with commodity switches and outperforms TCP by up to 60.5% if it bins the traffic into two priority queues according to flow size.
High Throughput and Low Latency on Hadoop Clusters Using Explicit Congestion Notification: The Untold Truth
TLDR
This paper studies the difficulty in configuration of TCP/IP extensions related to ACK packets, and explains how both TCP-ECN and DCTCP can achieve the best performance using a simple marking scheme, in constrast to the current preference for relying on AQMs to mark packets.
Support ECN in Multi-Queue Datacenter Networks via Per-Port Marking with Selective Blindness
  • Y. Pan, Chen Tian, Guihai Chen
  • Computer Science
    2018 IEEE 38th International Conference on Distributed Computing Systems (ICDCS)
  • 2018
TLDR
Per-Port Marking with Selective Blindness (PMSB) is presented and it is demonstrated that PMSB can preserve a given scheduling policy and reduce the average/99% completion time for small flows and a slightly better performance for large flows.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 47 REFERENCES
Fastpass: a centralized "zero-queue" datacenter network
TLDR
This paper describes Fastpass, a datacenter network architecture built using this principle that achieves high throughput comparable to current networks at a 240x reduction is queue lengths, and achieves much fairer and consistent flow throughputs than the baseline TCP.
Better never than late: meeting deadlines in datacenter networks
TLDR
Evaluation from a 19-node, two-tier datacenter testbed shows that D3, even without any deadline information, easily outperforms TCP in terms of short flow latency and burst tolerance, and effectively doubles the peak load that thedatacenter network can support.
Data center TCP (DCTCP)
Less Is More: Trading a Little Bandwidth for Ultra-Low Latency in the Data Center
TLDR
The HULL (High-bandwidth Ultra-Low Latency) architecture is presented to balance two seemingly contradictory goals: near baseline fabric latency and high bandwidth utilization and results show that by sacrificing a small amount of bandwidth, HULL can dramatically reduce average and tail latencies in the data center.
Reconciling high server utilization and sub-millisecond quality-of-service
TLDR
It is demonstrated that some latency-critical workloads can be aggressively co-located with other workloads, achieve good QoS, and that such co-location can improve a datacenter's effective throughput per TCO-$ by up to 52%.
Silo: Predictable Message Completion Time in the Cloud
TLDR
This work presents Silo, a network architecture for public cloud datacenters that offers guaranteed network bandwidth, guaranteed per-packet delay and guaranteed burst allowance and leverages the fact that guaranteed bandwidth and delay are tightly coupled.
Practical TDMA for datacenter ethernet
TLDR
A TDMA MAC layer for commodity Ethernet hardware is designed and implemented that allows end hosts to dispense with TCP's reliability and congestion control and is found to result in higher bandwidth for MapReduce shuffle workloads, lower latency for RPC workloads in the presence of background traffic, and more efficient operation in highly dynamic and hybrid optical/electrical networks.
Finishing flows quickly with preemptive scheduling
TLDR
It is demonstrated that PDQ significantly outperforms TCP, RCP and D3 in data center environments, and is stable, resilient to packet loss, and preserves nearly all its performance gains even given inaccurate flow information.
A generalized processor sharing approach to flow control in integrated services networks-the single node case
  • Abhay K. Parekh, R. Gallager
  • Computer Science, Business
    [Proceedings] IEEE INFOCOM '92: The Conference on Computer Communications
  • 1992
TLDR
The authors propose the use of a packet service discipline at the nodes of the network that is based on a multiplex scheme called generalized processor sharing (GPS) that is combined with leaky bucket rate admission control to provide flexible, efficient and fair use of the links.
Deadline-aware datacenter tcp (D2TCP)
TLDR
This work proposes Deadline-Aware Datacenter TCP (D2TCP), a novel transport protocol, which handles bursts, is deadline-aware, and is readily deployable and uses a novel congestion avoidance algorithm which uses ECN feedback and deadlines to modulate the congestion window via a gamma-correction function.
...
1
2
3
4
5
...