Programmable Packet Scheduling at Line Rate

Sivaraman, Anirudh; Subramanian, Sandya; Alizadeh, Mohammad; Chole, Sharad; Chuang, Shang-Tse; Agrawal, Anurag A.; Balakrishnan, Hari; Edsall, Tom; Katti, Sachin; McKeown, Nick

doi:10.1145/2934872.2934899

Cited by 190 publications

(95 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While today's switches do not support priority queues, recent work [59,60] has shown that it is feasible to implement programmable priority queues in emerging programmable switching chips [11]. This design allows packets to be inserted into the queue based on a programmable rank value that is computed before the packet is enqueued.…”

Section: The Numfabric Switchmentioning

confidence: 99%

See 1 more Smart Citation

NUMFabric

Nagaraj

Bharadia

Mao

et al. 2016

Proceedings of the 2016 ACM SIGCOMM Conference

Self Cite

View full text Add to dashboard Cite

We present NUMFabric, a novel transport design that provides flexible and fast bandwidth allocation control. NUMFabric is flexible: it enables operators to specify how bandwidth is allocated amongst contending flows to optimize for different service-level objectives such as weighted fairness, minimizing flow completion times, multipath resource pooling, prioritized bandwidth functions, etc. NUMFabric is also very fast: it converges to the specified allocation 2.3× faster than prior schemes. Underlying NUMFabric is a novel distributed algorithm for solving network utility maximization problems that exploits weighted fair queueing packet scheduling in the network to converge quickly. We evaluate NUMFabric using realistic datacenter topologies and highly dynamic workloads and show that it is able to provide flexibility and fast convergence in such stressful environments.

show abstract

Section: The Numfabric Switchmentioning

confidence: 99%

“…This design allows packets to be inserted into the queue based on a programmable rank value that is computed before the packet is enqueued. The paper [60] shows how STFQ can be realized with this design. We omit the details and refer the reader to [60].…”

Section: The Numfabric Switchmentioning

confidence: 99%

NUMFabric

Nagaraj

Bharadia

Mao

et al. 2016

Proceedings of the 2016 ACM SIGCOMM Conference

Self Cite

View full text Add to dashboard Cite

show abstract

“…There are also such as Intel's FlexPipe [6], Cavium's Xpliant [7] and Cisco's Doppler [8] and the other business, using similar programmable pipelines to improve data plane performance. In [9], they proposed a idea of programmable packet scheduler, in the switch to prepare the scheduling algorithm to determine the order of packet scheduling and queue priority.It achieved the programmable hardware, but the increase design cost. In [10] introduced a high-level language programming data plane algorithm, and these programs can be compiled on the emerging switch chip to run at low speed microcode.…”

Section: Research On Programmable Data Planementioning

confidence: 99%

Research on Programmable Data Plane Load Balancing based on Multipath Selection

He¹,

Ou²

2017

Proceedings of the 3rd Annual International Conference on Electronics, Electrical Engineering and Information Science (EEEIS 20

View full text Add to dashboard Cite

Abstract. The switching equipment on traditional networks have limited capabilities and the exchange of the switching equipment is often associated with long time frame and high development cost. SDN deployment controller under the flow table rules also exists to handle message delay. In this paper, we propose a scheme based on multipath selection of programmable data plane load balancing. We use the P4 language to directly write the data plane flow control multipath program, and then use the improved equivalent multipath routing path matching and port selection to achieve the path efficient allocation and traffic load balancing scheduling. Finally, the NS2 simulation results show that, compared with the traditional equivalent multipath, the improved multipath selection scheme can lower the end-to-end delay, reduce the packet loss rate and improve the throughput.

show abstract

“…These algorithms encompass data-plane traffic engineering, in-network congestion control, active queue management, network security, and measurement. We also used Domino to express the priority computation for programming scheduling using push-in first-out queues [58].…”

Section: Expressivenessmentioning

confidence: 99%

“…These algorithms process and transform packets, reading and writing state in the switch. Examples include active queue management [38,47,51], scheduling [58], congestion control with switch feedback [45,60], network measurement [63,37], and data-plane traffic engineering [21].…”

Section: Introductionmentioning

confidence: 99%

Packet Transactions

Sivaraman¹,

Cheung

Budiu

et al. 2016

Proceedings of the 2016 ACM SIGCOMM Conference

Self Cite

208

View full text Add to dashboard Cite

Many algorithms for congestion control, scheduling, network measurement, active queue management, and traffic engineering require custom processing of packets in the data plane of a network switch. To run at line rate, these dataplane algorithms must be implemented in hardware. With today's switch hardware, algorithms cannot be changed, nor new algorithms installed, after a switch has been built. This paper shows how to program data-plane algorithms in a high-level language and compile those programs into low-level microcode that can run on emerging programmable line-rate switching chips. The key challenge is that many data-plane algorithms create and modify algorithmic state. To achieve line-rate programmability for stateful algorithms, we introduce the notion of a packet transaction: a sequential packet-processing code block that is atomic and isolated from other such code blocks.We have developed this idea in Domino, a C-like imperative language to express data-plane algorithms. We show with many examples that Domino provides a convenient way to express sophisticated data-plane algorithms, and show that these algorithms can be run at line rate with modest estimated chip-area overhead.

show abstract

Programmable Packet Scheduling at Line Rate

Cited by 190 publications

References 33 publications

NUMFabric

NUMFabric

Research on Programmable Data Plane Load Balancing based on Multipath Selection

Packet Transactions

Contact Info

Product

Resources

About