AppTCP: The design and evaluation of application-based TCP for e-VLBI in fast long distance networks
Wang, G; Wu, Y; Dou, K; et al.Ren, Y; Li, J
Date: 1 October 2014
Journal
Future Generation Computer Systems
Publisher
Elsevier
Publisher DOI
Abstract
Electric Very Long Baseline Interferometry (e-VLBI) is a typical astronomical interferometry used in radio astronomy. It allows
observations of an object that are made simultaneously by many radio telescopes to be combined, emulating a telescope with the size equal to the maximum separation between the telescopes. The main requirements ...
Electric Very Long Baseline Interferometry (e-VLBI) is a typical astronomical interferometry used in radio astronomy. It allows
observations of an object that are made simultaneously by many radio telescopes to be combined, emulating a telescope with the size equal to the maximum separation between the telescopes. The main requirements of transporting e-VLBI data are the high and constant transmission rate. However, the traditional TCP and its variants cannot meet these requirements. In an effort to solve the problem of transporting e-VLBI data in fast long distance networks, we propose an application-based TCP (AppTCP) congestion control algorithm, using Closed-Loop Control theory to keep the stable and constant transmission rate. AppTCP can swiftly reach the required transmission rate by increasing the congestion control window, and keep the transmission rate and allows the other TCP flows to share the remaining bandwidth. We further conduct extensive experiments in both fast long distance network test-bed and actual national networks (i.e., from Beijing to Shanghai in China) and international networks (i.e., from Hongkong in China to Chicago in USA) to evaluate and verify the performance and effectiveness of AppTCP. The results show that the AppTCP can effectively utilize the link capacity and maintain the constant rate during the data transmission, and its performance significantly outperforms that of the existing TCP variants.
Mathematics and Statistics
Faculty of Environment, Science and Economy
Item views 0
Full item downloads 0