Low-Latency Scheduling in MPTCP

Hurtig, Per; Grinnemo, Karl-Johan; Brunström, Anna; Ferlin, Simone; Alay, Özgü; Kuhn, Nicolas

doi:10.1109/tnet.2018.2884791

Cited by 101 publications

(60 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To solve the receiver buffer problem [32] in multipath transport, [30] proposes the Penalisation and Retransmission (PR) mechanism. [28] derives a rule-of-thumb for buffer size for MPTCP based on the RTT difference of the paths; and [29] explores an implementation that could send packets outof-order, both with the ambition to overcome head-of-line (HoL) blocking.…”

Section: B the State Of The Art Schedulersmentioning

confidence: 99%

“…More recently, [37] proposes a lossaware scheduler targeting networks with more than 20% loss rates. Finally, [32] proposes the Short Transfer Time First (STTF) scheduler. However, it specifically targets low-latency for short transfers, and it explores interactions with TCP specific aspects such as TCP Small Queues (TSQ).…”

Section: B the State Of The Art Schedulersmentioning

confidence: 99%

See 1 more Smart Citation

Peekaboo: Learning-Based Multipath Scheduling for Dynamic Heterogeneous Environments

Alay

Brunström

et al. 2020

IEEE J. Select. Areas Commun.

Self Cite

View full text Add to dashboard Cite

Multipath transport protocols utilize multiple network paths (e.g., WiFi and cellular) to achieve improved performance and reliability, compared with their single-path counterparts. The scheduler of a multipath transport protocol determines how to distribute the data packets onto different paths. However, state-of-the-art multipath schedulers face the challenge when dealing with heterogeneous paths with dynamic path characteristics (i.e., packet loss, fluctuation of delay). In this paper, we propose Peekaboo, a novel learning-based multipath scheduler that is aware of the dynamic characteristics of the heterogeneous paths. Peekaboo is able to learn scheduling decisions to adopt over time based on the current path characteristics and dynamicity levels -from both deterministic and stochastic perspectives. We implement Peekaboo in Multipath QUIC (MPQUIC) and compare it with state-of-the-art multipath schedulers for a wide range of dynamic heterogeneous environments, upon both emulated and real networks. Our results show that Peekaboo outperforms the other schedulers by up to 31.2% in emulated networks and up to 36.3% in real network scenarios.

show abstract

Section: B the State Of The Art Schedulersmentioning

confidence: 99%

Section: B the State Of The Art Schedulersmentioning

confidence: 99%

Peekaboo: Learning-Based Multipath Scheduling for Dynamic Heterogeneous Environments

Alay

Brunström

et al. 2020

IEEE J. Select. Areas Commun.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In order to avoid the asymmetry problem caused by the latency difference, STMS [37] scheduler sends packets with smaller sequence number to a fast path while sending packets with larger sequence number to a slow path, allowing packets sent over different paths to arrive simultaneously. Considering the continuously increasing demand for mobile communication, literature [38] presents two novel scheduling schemes which are the block estimation (BLEST) scheduler and the shortest transmission time first (STTF) scheduler to guarantee low-latency communication of MPTCP. To improve the performance of latency-sensitive applications over MPTCP in high-delay and lossy networks, literature [39] proposes a new framework using a XOR-based dynamic FEC scheme to reduce the flow completion time.…”

Section: Related Workmentioning

confidence: 99%

Tuning high flow concurrency for MPTCP in data center networks

Huang

et al. 2020

J Cloud Comp

View full text Add to dashboard Cite

In the data center networks, multipath transmission control protocol(MPTCP) uses multiple subflows to balance traffic over parallel paths and achieve high throughput. Despite much recent progress in improving MPTCP performance in data center, how to adjust the number of subflows according to network status has remained elusive. In this paper, we reveal theoretically and empirically that controlling the number of concurrent subflows is very important in reducing flow completion time (FCT) under network dynamic. We further propose a novel design called MPTCP_OPN, which adaptively adjusts the number of concurrent subflows according to the real-time network state and flexibly shifts traffic from congested paths to mitigate the high tail latency. Experimental results show that MPTCP_OPN effectively reduces the timeout probability caused by full window loss and flow completion time by up to 50% compared with MPTCP protocol.

show abstract

“…Coding: Network Coding Scheduling: Paths Hurtig et al [103] Scheduling: Paths QoS-MPTPC [104] Scheduling: Packets ADMIT [105] Coding: FEC…”

Section: Fmtcp [102]mentioning

confidence: 99%

“…The NC-MPTCP [101] and fountain code-based multipath TCP (FMTCP) [102] utilize network coding to boost the overall goodput and outperform the MPTCP in the case of highly dissimilar subflow conditions. Hurtig et al [103] present two novel scheduling techniques for the MPTCP (BLEST and STTF) that are shown to reduce latency when interfaces have asymmetric capacity and delay. The QoS-MPTPC [104], ADMIT [105] and PR-MPTCP+ [106] are extensions for interactive video, video streaming and real-time multimedia, respectively.…”

Section: ) Multiple Connectivity and Multipath Protocolsmentioning

confidence: 99%

A Survey of End-to-End Solutions for Reliable Low-Latency Communications in 5G Networks

Rico

Merino

2020

IEEE Access

View full text Add to dashboard Cite

Very low latency and high reliability are two of the main requirements for new applications exploiting 5G networks. This is the case for the remote operation of robots or vehicles, the autonomous interaction of equipment in a factory, autonomous driving and tactile internet applications. Although the TCP/IP stack has been sufficient as the end-to-end solution for most of the history of the Internet, a number of surveys have appeared recently presenting many different methods for managing the end-to-end communication to meet the requirements of various technologies such as that of 5G networks. In this paper, we present a novel classification of the literature focused on new end-to-end solutions and the creation of services towards the support of low latency (1 ms) and high reliability (10-9 error rate) in current and future 5G networks. We specifically highlight how the proposals can be classified according to enabling technologies and the specific method used to achieve success in terms of the latency and reliability. The literature related to end-to-end solutions for reliable low-latency communications are organized according to three main topics: (i) end-to-end protocols that improve communication in terms of latency and reliability, (ii) functionality or technologies implemented on the network to support the current demands, and (iii) application programming interfaces that enhance the correct utilization of those protocols and additional technologies.

show abstract

Low-Latency Scheduling in MPTCP

Cited by 101 publications

References 41 publications

Peekaboo: Learning-Based Multipath Scheduling for Dynamic Heterogeneous Environments

Peekaboo: Learning-Based Multipath Scheduling for Dynamic Heterogeneous Environments

Tuning high flow concurrency for MPTCP in data center networks

A Survey of End-to-End Solutions for Reliable Low-Latency Communications in 5G Networks

Contact Info

Product

Resources

About