Experimental Evaluations of Diffusion-Based Autonomous Decentralized Congestion Control

Tatsuro SUMI  Chisa TAKANO  Masaki AIDA  Kenji ISHIDA  

D - Abstracts of IEICE TRANSACTIONS on Information and Systems (Japanese Edition)   Vol.J95-D   No.12   pp.2048-2058
Publication Date: 2012/12/01
Online ISSN: 1881-0225
Print ISSN: 1880-4535
Type of Manuscript: PAPER
autonomous decentralized control,  diffusion equation,  traffic control,  congestion control,  experimental evaluation,  

Full Text(in Japanese): PDF(1.1MB)
>>Buy this Article

Various applications, such as video streaming service and social networking service, have appeared in recent years, and use of new mobile terminals such as a smart phone is increasing. Then, network state becomes unpredictable by interaction of network state and users behavior, while network traffic is increasing explosively. Therefore, a high-speed network requires flexible and powerful network control which guarantees the wide variety of application requirements. Generally, in a high-speed network, it is impossible to implement time-sensitive control based on collecting global information about the whole network. If we allow sufficient time to collect global information, the gathered data includes old information, and so the information cannot be used for time-sensitive control. Our previous studies proposed diffusion type flow control (DFC) as a solution for the extremely time-sensitive flow control required by high-speed networks and we have evaluated the performance of DFC by simulation. In this paper, to apply DFC to real networks, we show the technique to implement the DFC functions on PCs and investigate its performance.