DocumentCode
267099
Title
Modeling and Understanding TCP´s Fairness Problem in Data Center Networks
Author
Shuli Zhang ; Yan Zhang ; Yifang Qin ; Yanni Han ; Song Ci
Author_Institution
High Performance Network Lab., Inst. of Acoust., Beijing, China
fYear
2014
fDate
15-18 Dec. 2014
Firstpage
605
Lastpage
610
Abstract
Due to the special topologies and communication pattern, in today´s data center networks it is common that a large set of TCP flows and a small set of TCP flows get into different ingress ports of a switch and compete for a same egress port. However, in this case the throughput share of flows in the two sets will not be fair even though all flows have the same RTT. In this paper, we study this problem and find that TCP´s fairness in data center networks is related with not only the network capacity but also the number of flows in the two sets. We propose a mathematical model of the average throughput ratio of the large set of flows to the small set of flows. This model can reveal the variation of TCP´s fairness along with the change of network parameters (including buffer size, bandwidth, and propagation delay) as well as the number of flows in the two sets. We validate our model by comparing its numerical results with simulation results, finding that they match well.
Keywords
computer centres; telecommunication network topology; transport protocols; TCP fairness problem modeling; TCP fairness problem understanding; TCP flows; average throughput ratio mathematical model; bandwidth; buffer size; communication pattern; data center networks; egress port; propagation delay; switch ingress ports; topology pattern; Bandwidth; Ports (Computers); Propagation delay; Steady-state; Switches; Throughput; Topology; TCP´s fairness; data center networks; model; network capacity;
fLanguage
English
Publisher
ieee
Conference_Titel
Cloud Computing Technology and Science (CloudCom), 2014 IEEE 6th International Conference on
Conference_Location
Singapore
Type
conf
DOI
10.1109/CloudCom.2014.153
Filename
7037723
Link To Document