TCP congestion control in data center networks
dc.contributor.author | Tahiliani, R.P. | |
dc.contributor.author | Tahiliani, M.P. | |
dc.contributor.author | Chandra Sekaran K. | |
dc.date.accessioned | 2020-03-31T14:15:19Z | |
dc.date.available | 2020-03-31T14:15:19Z | |
dc.date.issued | 2015 | |
dc.description.abstract | Data Center Networks (DCNs) provide the core infrastructure backbone for various small and large scale enterprise applications, including Cloud Computing applications. TCP is used by a plethora of Internet applications today and it also retains its dominance in DCNs, in terms of the most widely deployed transport protocol. The nature of DCN traffic, however, largely differs from that of the Internet. The fundamental assumptions made during the initial design of TCP hinder its suitability in state-of-the-art DCNs. Impairments such as TCP Incast, TCP Outcast, Buffer pressure, Queue build-up and pseudo-congestion effect severely degrade the performance of traditional TCP variants in DCNs. With cloud computing becoming an important part of the forseeable future, it is extremely crucial to eliminate these impairments. First, this chapter describes the above mentioned impairments in brief and explores TCP variants proposed so far to overcome the same. Second, the advantages and shortcomings of each TCP variant are highlighted in terms of its effificacy and the deployment complexity. Open research issues related to the TCP’s performance in DCNs are also discussed. © Springer Science+Business Media New York 2015. | en_US |
dc.identifier.citation | Handbook on Data Centers, 2015, Vol., pp.485-505 | en_US |
dc.identifier.uri | 10.1007/978-1-4939-2092-1_15 | |
dc.identifier.uri | https://idr.nitk.ac.in/jspui/handle/123456789/13731 | |
dc.title | TCP congestion control in data center networks | en_US |
dc.type | Book Chapter | en_US |