TCP Variants for Data Center Networks: A Comparative Study

Tahiliani, Rohit P.; Tahiliani, Mohit P.; Sekaran, K. Chandra

doi:10.1109/iscos.2012.38

Cited by 16 publications

(15 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Under this traffic pattern, the higher-level aggregators receive the data request from the client and then partition into several pieces and transfer to the workers via the lower-level aggregators. According to [5,14,18], the data center traffic can be categorized into three types:…”

Section: Data Center Traffic and Tcp Incastmentioning

confidence: 99%

“…Data center networks are finely designed and layered to achieve high bandwidth and low latency. As a result, the data center environment is highly different than that of the Internet, especially in terms of round trip time (RTT) [5]. The main characteristics of a data center network are high-speed links, limited-size switch buffers, low propagation delays and fault tolerance.…”

Section: Introductionmentioning

confidence: 99%

“…Data center traffic can be classified into three types, namely elephant traffic, mice traffic and cat traffic [5]. According to recent studies [5][6][7], above 90 % of network traffic in a data center is contributed by transmission control protocol (TCP) due to its reliability of data transmission.…”

Section: Introductionmentioning

confidence: 99%

“…According to recent studies [5][6][7], above 90 % of network traffic in a data center is contributed by transmission control protocol (TCP) due to its reliability of data transmission. In data center network, there are three basic requirements for the data center transport such as high throughput, low latency and high burst tolerance.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

An early congestion feedback and rate adjustment schemes for many-to-one communication in cloud-based data center networks

2015

View full text Add to dashboard Cite

Cloud data centers are playing an important role for providing many online services such as web search, cloud computing and back-end computations such as MapReduce and BigTable. In data center network, there are three basic requirements for the data center transport protocol such as high throughput, low latency and high burst tolerance. Unfortunately, conventional TCP protocols are unable to meet the requirements of data center transport protocol. One of the main practical issues of great importance is TCP Incast to occur many-to-one communication sessions in data centers, in which TCP experiences sharp degradation of throughput and higher delay. This important issue in data center networks has already attracted the researchers because of the development of cloud computing. Recently, few solutions have been proposed for improving the performance of TCP in data center networks. Among that, DCTCP is the most popular protocol in academic as well as industry areas due to its better performance in terms of throughput and latency. Although DCTCP provides significant performance improvements, there are still some defects in maintaining the queue length and throughput when the number of servers is too large. To address this problem, we propose a simple and efficient TCP protocol, namely NewDCTCP as an improvement of DCTCP in data center networks. NewDCTCP modified the conges-B Meejeong Lee tion feedback and window adjusting schemes of DCTCP to mitigate the TCP Incast problem. Through detailed QualNet experiments, we show that NewDCTCP significantly outperforms DCTCP and TCP in terms of goodput and latency. The experimental results also demonstrate that NewDCTCP flows provide better link efficiency and fairness with respect to DCTCP.

show abstract

Section: Data Center Traffic and Tcp Incastmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

An early congestion feedback and rate adjustment schemes for many-to-one communication in cloud-based data center networks

2015

View full text Add to dashboard Cite

show abstract

“…There are three types of traffic that exist in DCNs : mice traffic or query traffic (small query response traffic like web search, facebook notification etc), cat traffic (medium size downloads and photo uploads etc) and elephant traffic (software downloads, large video downloads) [4]. Majority of the traffic in DCNs is usually query traffic and carries less data volume.…”

Section: Introductionmentioning

confidence: 99%

Revisiting TCP Outcast Problem Using DCTCP in Data Center Networks

Patidar

Tahiliani

2015

2015 Fifth International Conference on Advanced Computing &Amp; Communication Technologies

Self Cite

View full text Add to dashboard Cite

Data Center Networks (DCNs) today are used to compute and store the data of multiple users because of low cost, flexibility and availability. In this multi-tenant cloud environment, resources are shared among different users. Suppose two users fetch data at same time from the same DCNs, one fetches small amount of data (like web search) while another user fetches larger amount of data (like software update). If these large set of flows and small set of flows arrive at two different input ports of a switch and compete for same output port, then large flow packets get the buffer while small flow packets are dropped consecutively. This issue occurs due to droptail implementation of queue in switches and as one set of flows is outcast by other, the problem is known as TCP Outcast. While a few studies have already explored the TCP Outcast problem, this paper aims to revisit this problem by using DCTCP in DCNs. The experiments carried out using Mininet show that DCTCP solves the problem of TCP Outcast by ensuring fairness for all types of flows.

show abstract