Loading…
An adaptive data chunk scheduling for concurrent multipath transfer
Concurrent Multipath Transfer (CMT) is a transport layer protocol which provides concurrent data transfer over the multiple paths. CMT improves the available bandwidth utilization, fault tolerance, robustness and reliability of the network. However, in multipath data transfer, destination receives o...
Saved in:
Published in: | Computer standards and interfaces 2017-05, Vol.52, p.97-104 |
---|---|
Main Authors: | , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Concurrent Multipath Transfer (CMT) is a transport layer protocol which provides concurrent data transfer over the multiple paths. CMT improves the available bandwidth utilization, fault tolerance, robustness and reliability of the network. However, in multipath data transfer, destination receives out-of-order data chunk due to dissimilar delay and bandwidth of each path. It causes the serious problem of receiver buffer blocking, unwanted congestion window (cwnd) reduction and unnecessary retransmission, which significantly degrades the performance of CMT. Thus, this paper proposes an adaptive data chunk scheduling for CMT (A-CMT). The proposed method uses path delay and bandwidth as a factor of data chunk scheduling to adapt path conditions. The simulation results show that proposed method achieves better performance in terms of throughput; file transfer time and congestion window growth. The proposed method improves average throughput up to 13%.
•We introduce an adaptive data chunk scheduling policy for CMT-SCTP.•The proposed policy uses the delay and bandwidth of the path as factor to identify the path status.•We use RTT (Round Trip Time) and cwnd (congestion window) of path to estimate path quality.•Simulation results show better throughput, improved congestion window growth and less file transfer time. |
---|---|
ISSN: | 0920-5489 1872-7018 |
DOI: | 10.1016/j.csi.2017.02.001 |