BLEST: Blocking estimation-based MPTCP scheduler for heterogeneous networks

Ferlin, Simone; Alay, Özgü; Mehani, Olivier; Boreli, Roksana

doi:10.1109/ifipnetworking.2016.7497206

Cited by 142 publications

(100 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We utilize our Kernel implementation summarized in Section V. The experiments were performed in a university data center and consider a variety of workloads such as video streaming, web file downloads, etc. We compare QAware with several state-of-the-art schedulers such as min-SRTT, Delay Aware Packet Scheduler (DAPS) [15], Blocking Estimation based scheduler (BLEST) [8], and ECF [17]. The details of our experiments and consequent results are discussed in Section VIII.…”

Section: Evaluation Methodologymentioning

confidence: 99%

“…However, this makes DAPS unable to react promptly to network changes due to pre-computed long schedules. Blocking-Estimation-based MPTCP Scheduler (BLEST) [8] aims to reduce head-of-line blocking by waiting for the faster subflow despite the availability of space in congestion window of the slower subflow. ECF [17] follows a similar principle as that of BLEST, but while BLEST aims to reduce out-of-order delivery assuming that the send buffer is a bottleneck, ECF aims to minimize completion time.…”

Section: Background and Related Workmentioning

confidence: 99%

“…We compare QAware with the following schedulers: i) min-SRTT, ii) Delay Aware Packet Scheduler (DAPS) [15] iii) Blocking Estimation based Scheduler (BLEST) [8], and iv) Earliest Completion First (ECF) [17] 7 8 . We first compare scheduler performance for application generating bulk traffic.…”

Section: Real-world Setup and Experimentsmentioning

confidence: 99%

“…Scheduling between the multiple connections is an obvious research problem and recently multiple proposals [8], [9], [15], [17] have emerged to improve the default MPTCP scheduler [20]. Typically, these schedulers use a transport layer estimate of the end-to-end bandwidth/delay (for example, the smoothed round-trip time) for each TCP subflow as an input to the scheduling policy that decides on how application data must be assigned to the multiple subflows.…”

Section: Introductionmentioning

confidence: 99%

“…This enables us to leverage queueing theoretical ISBN 978-3-903176-08-9 c 2018 IFIP insights to create a scheduling policy that combines endto-end delays and device driver queue occupancy. (c) Our simulations and real-world experimentation over a wide range of applications compare QAware with the default MPTCP scheduler [20], ECF [17], DAPS [15], and BLEST [8].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

QAware: A Cross-Layer Approach to MPTCP Scheduling

Shreedhar

Mohan

Kaul

et al. 2018

2018 IFIP Networking Conference (IFIP Networking) and Workshops

View full text Add to dashboard Cite

Multipath TCP (MPTCP) allows applications to transparently use all available network interfaces by creating a TCP subflow per interface. One critical component of MPTCP is the scheduler that decides which subflow to use for each packet. Existing schedulers typically use estimates of end-to-end path properties, such as delay and bandwidth, for making the scheduling decisions. In this paper, we show that these scheduling decisions can be significantly improved by including readily available local information from the device driver queues to the decision-making process. We propose QAware, a novel cross-layer approach for MPTCP scheduling. QAware combines end-to-end delay estimates with local queue buffer occupancy information and allows for a better and faster adaptation to the network conditions. This results in more efficient use of the available resources and considerable gains in aggregate throughput. We present the design of QAware and evaluate its performance through simulations, and also through real experiments, comparing it to existing schedulers. Our results show that QAware performs significantly better than other available approaches for all use-cases and applications.

show abstract

Section: Evaluation Methodologymentioning

confidence: 99%

Section: Background and Related Workmentioning

confidence: 99%

Section: Real-world Setup and Experimentsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

QAware: A Cross-Layer Approach to MPTCP Scheduling

Shreedhar

Mohan

Kaul

et al. 2018

2018 IFIP Networking Conference (IFIP Networking) and Workshops

View full text Add to dashboard Cite

show abstract

A reliable alternating multi‐path transmission control protocol scheduling for bandwidth aggregation performance

Bencheikh

Zouaoui

Taleb

et al. 2023

Concurrency and Computation

View full text Add to dashboard Cite

SummaryThe Multi‐Path Transmission Control Protocol is a new generation protocol for the transport layer which allows multihomed devices to exploit simultaneously their own interfaces. This feature would normally lead to some high‐performance achievement in regard to regular TCP. However, the heterogeneous paths might cause the occurrence of the 'out‐of‐order' problem which requires the intervention of a more efficient scheduler to reach a high level of global bandwidth utilization. Since scheduling plays a key role in performance improvements, several studies have been implemented exploiting this idea, namely those that adopt the reactive approach and those others adopting the proactive one. In this paper, our goal is to improve the multipath throughput. The testbed experimentation consists primarily in the evaluation and analysis of the behaviour of different recent scheduling approaches with an asymmetric paths topology. Then, we propose an alternating scheduling approach which shed the light on bandwidth utilization. The results show that this approach is a more efficient solution than taking the reactive and proactive ones separately as it is based on a preliminary study that we conducted on the main performance and reliability parameters and has been implemented in the asymmetric multipath networks. This technique has led to a smart scheduling that provides a high level of throughput aggregation.

show abstract

Device‐centric data reordering and buffer management for mobile Internet using Multipath Transmission Control Protocol

Kumar

Tyagi

2021

Int J Communication

View full text Add to dashboard Cite

Summary Today's smart mobile devices (SMDs) are equipped with multiple interfaces, powerful processors, and workable battery power. Such SMDs are making future networks truly heterogeneous. In such future mobile networks, there is an opportunity to utilize the SMD's interfaces concurrently. There are challenges in the implementation of this idea, because regular TCP does not support path aggregation, multihoming, and full‐mesh networking. The multipath TCP (MPTCP) is a power set of TCP extensions that provide the subflows on each SMD's interface. The MPTCP runs over the existing infrastructure of the TCP by using TCP options field, without compromising the normal TCP functions. An end‐to‐end data scheduling and data reordering are important issues in such mobile Internet across multiple subflows of MPTCP. In this paper, we proposed the mechanisms for multipath scheduling, device‐centric data reordering, and buffer management to achieve concurrent utilization of SMD's LTE and Wi‐Fi interfaces for video stream applications on mobile Internet. We modeled the approaches for multipath scheduling and reordering to ensure the improved utilization of SMD's interfaces and decrease the buffer usage. The proposed approaches consider the outstanding data bytes and latency of each MPTCP subflow. The proposed approaches are tested on a MPTCP‐supported testbed, and captured data are analyzed in terms of gross throughput, buffer usage, end‐to‐end performance, and path utilization of subflows on SMD's LTE and Wi‐Fi interfaces. The performance is evaluated and compared with related state‐of‐the‐art approaches, and significant improvement is observed.

show abstract

BLEST: Blocking estimation-based MPTCP scheduler for heterogeneous networks

Cited by 142 publications

References 13 publications

QAware: A Cross-Layer Approach to MPTCP Scheduling

QAware: A Cross-Layer Approach to MPTCP Scheduling

A reliable alternating multi‐path transmission control protocol scheduling for bandwidth aggregation performance

Device‐centric data reordering and buffer management for mobile Internet using Multipath Transmission Control Protocol

Contact Info

Product

Resources

About