Performance evaluation of TCP congestion control algorithms throughput for CVE based on cloud computing model

Abstract

Collaborative Virtual Environment (CVE) is becoming popular in the last few years; this is because CVE is designed to allow geographically distributed users to work together over the network. Currently, in the development of CVE Systems, Client server architectures with multiple servers are used with TCP as update transmitting transport protocol because of its reliability. With the increasing number of collaborators, the transport protocol is inadequate to meet the system requirements in terms of timely data transmission. The transport protocol (TCP) throughput deteriorates in the network with large delay which leads to unsatisfactory consistency requirement of the CVE systems.We proposed a cloud based architectural model for improving scalability and consistency in CVE in an earlier study. The current paper aims at evaluating and comparing the performance of different TCP variants (Tahoe, Reno, New Reno, Vegas, SACK, Fack and Linux) with the cloud based CVE architecture to determine the suitability of each TCP variant for CVE. A comparative analysis between the different TCP variants is presented in terms of throughput verses elapse time, with increasing number of users in the system. TCP Vegas with the cloud based model was found to be effective for CVE systems based on Cloud Computing