P4QCN: Congestion Control Using P4-Capable Device in Data Center Networks

Abstract

Modern data centers aim to offer very high throughput and ultra-low latency to meet the demands of applications such as online intensive services. Traditional TCP/IP stacks cannot meet these requirements due to their high CPU overhead and high-latency. Remote Direct Memory Access (RDMA) is an approach that can be designed to meet this demand. The mainstream transport protocol of RDMA over Ethernet is RoCE (RDMA over Converged Ethernet), which relies on Priority Flow Control (PFC) within the network to enable a lossless network. However, PFC is a coarse-grained protocol which can lead to problems such as congestion spreading, head-of-the-line blocking. A congestion control protocol that can alleviate these problems of PFC is needed. We propose a protocol, called P4QCN for this purpose. P4QCN is a congestion control scheme for RoCE and it is an improved Quantized Congestion Notification (QCN) design based on P4, which is a flow-level, rate-based congestion control mechanism. P4QCN extends the QCN protocol to make it compatible with IP-routed networks based on a framework of P4 and adopts a two-point algorithm architecture which is more effective than the three-point architecture used in QCN and Data Center QCN(DCQCN). Experiments show that our proposed P4QCN algorithm achieves the expected performance in terms of latency and throughput.

An Exhaustive Survey on P4 Programmable Data Plane Switches: Taxonomy, Applications, Challenges, and Future Trends

“…• Reconfigurability: the parser and the processing logic can be redefined in the field. Variations [54][55][56][57][58][59][60][61][62] Collectors and Solutions [63][64][65][66][67] Congestion Control [68][69][70][71][72][73][74][75][76] Measurements AQM [99][100][101][102][103][104][105][106][107][108][109] QoS and TM [110][111][112][113][114] Multicast [115][116][117] Load Balancing [118][119][120][121][122][123][124][125]…”

“…The goal of FAB is to minimize the flow completion time for short flows in specific workloads. Geng et al [76] proposed P4QCN, a flow-level, rate-based congestion control mechanism that improves the Quantized Congestion Notification (QCN). P4QCN improves QCN by alleviating the problems of PFC within a lossless network.…”

A Review of P4 Programmable Data Planes for Network Security

“…Engineering. With the emergence of programmable switches, fine-grained measurements can be performed at line rate, so functions such as congestion detection [19], active queue management [20], and load balancing [21] can be realized to improve the quality of service.…”

Data Center TCP com Histerese em Switches P4

“…Um trabalhos relacionado importante é o P4qcn [Geng et al 2019]. Nele, foi criado um par de algoritmos capazes de tratar o congestionamento na rede.…”

“…O uso de P4 para solucionar problemas de congestionamento em switches não é uma novidade em si. Como, por exemplo, em [Geng et al 2019], os autores construíram um algoritmo que prevê congestionamento e gera pacotes de realimentac ¸ão que informam a necessidade da reduc ¸ão na taxa de envio. Eles utilizam apenas um limiar para conferir se ocorreu congestionamento na rede e, ao invés de usarem o DCTCP para a reduc ¸ão, operam com um algoritmo próprio para cumprir essa tarefa e pressupõe que a camada de enlace seja compatível com RoCE (Remote Direct Memory Access -RDMA) sobre Converged Ethernet.…”