Multipath transport and packet spraying for efficient data delivery in data centres

Abstract

Modern data centres provide large aggregate network capacity and multiple paths among servers. Traffic in data centres is very diverse; most of the data is produced by long, bandwidth hungry flows but the large majority of flows, which commonly come with stringent deadlines regarding their completion time, are short. It has been shown that TCP is not efficient for any of these types of traffic in modern data centres. MultiPath TCP (MPTCP) employs multipath data transport and is efficient for long flows but ill-suited for short flows. In this paper, we present Maximum MultiPath TCP (MMPTCP), a novel transport protocol which extends MPTCP and, compared to TCP and MPTCP, reduces short flows' completion times, while providing excellent goodput to long flows. To do so, MMPTCP runs in two phases; initially, it randomly scatters packets in the network under a single congestion window exploiting all available paths. This is beneficial to latency-sensitive flows. After a specific amount of data is sent, MMPTCP switches to a regular MultiPath TCP mode. MMPTCP is incrementally deployable in existing data centres as it does not require any modifications outside the transport layer and behaves well when competing with MPTCP flows. We also present a topology-specific extension of MMPTCP that adjusts the numbers of subflows during the second phase of the protocol based on knowledge about the location of the receiver in the data centre. We present extensive evaluation that shows that MMPTCP's design objectives are met. We have implemented MMPTCP (along with MPTCP and packet spraying) in ns-3 and evaluated our protocol in simulated FatTree topologies. We have evaluated how MMPTCP performs compared to TCP and MPTCP and how its performance is affected by transient hotspots in the network. We have also experimented with different thresholds for duplicate acknowledgements and fast retransmissions and shown that MMPTCP performs well when the size of short flows is widely ranged. Finally, we have evaluated how MMPTCP performs under conditions that result in Incast, when different congestion control algorithms are used in its second phase and when varying the overall network load.

Utility-Based Multipath Delivery of Prioritized XR Content in a Machine Learning and Network Slicing-enhanced Environment

“…An open-source default MPTCP model [17] based on NS-3 was developed adhering to the principles of transparency and TCP backward compatibility outlined in IETF RFC 6824. The solution proposed in our paper extends upon this NS-3based implementation.…”

CMT-SCTP and MPTCP Multipath Transport Protocols: A Comprehensive Review

“…[108,141,199 The scheme maintains the packet assignment ratio to the two subflows equivalent to the throughput ratio of the considered two subflows. Oh and Lee [211] CL-ADSP [146] A-DSP [147] Peng et al [133] Ferlin et al [135] Kaiping et al [168] Zhu et al [219] Elgabli et al [220] Konsgen et al [221] MPTCP [8,122] Xu et al [142] Wu et al [117] NC-MPTCP [48] Ni et al [160] Yang and Amer [159] Lim et al [151] A-MPTCP [209,210] Ferlin et al [217] Thomas et al [226] Pang et al [222] Dong et al [60] CWA-MPTCP [144] Le and Bui [158] Cao et al [212] Choi et al [195] AMTCP [209] Jiyan et al [121] Mmtcp [213] Cui et al [214] Wang et al [172] Kimaura et al [202] BELIA [38] Wu et al [217] Trinh et al [174] Li et al [223] Electronics…”

Survey on Traffic Management in Data Center Network: From Link Layer to Application Layer

“…This can degrade the performance of latency-sensitive short flows. To cope with this problem, MMPTCP [50] distinguishes the long and short flows based on the number of bytes have been sent and employs different transmission schemes for long and short flows. For the short flows, packet granularity is used to seize available path resources and reduce flow completion time quickly.…”