Font Size: a A A

Research On Flow-state Aware Network Detection And Traffic Control Based On IPMAN

Posted on:2007-09-01Degree:MasterType:Thesis
Country:ChinaCandidate:J HuaFull Text:PDF
GTID:2178360185497210Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Flow-state Aware Network Detection and Traffic Control Based on IPMAN is an innovational kind of technology in Traffic Engineering domain. Along with the P2P application affecting Network Design,the quality of Internet Data transmission will not be guaranteed and Data business cannot be estimated. The control mechanism and behavior characteristics become more and more complex. Though traditional QOS of IP network can provide some SLA with Diffserv, it cannot satisfy the need of macroscopical control of IPMAN for ISP. In this paper, with the analysis of P2P traffic behavior characteristics, we studied an innovational kind of flow state aware network detection and traffic fuzzy congestion control based on the IP network architecture. Combined with the traditional QOS technology, it provides multi-layers customization service in order to satisfy the need of traffic manager for the control of 'IP ServiceIn this paper, we make a profound study in the field of IPMAN Traffic Status and deploy the technology in the practical QOS Project. Firstly, we analyse traditional QOS technology(IntServ model and Diffserv model),which only focused on source/destination address,port number and protocol ,without taking into account Transmission Byte ,average Rate , Transfer Duration, Packet Size and etc. 78% of current Internet Data Flow is P2P application , such as BT ,Skype , and etc. Because of the dynamic distribution of Port,the traditional QOS cannot resolve the problem of network detection and bandwidth management. The different service type can be determined and marked through analysis of the characteristics of P2P Application model based on IPMAN. Then, one kind of Flow-state Aware Network Detection is designed. The Service Control architecture must fit the detected Data Flow. In order to overcome the limitation of Markov model, referred to the self-seminal characteristic of IPMAN ,we make fuzzy congestion control architecture ,which satisfy the need of IP Business characteristics and can be realized easily. This kind of technology is deployed in Shanghai Telecom Network,an ISP's Level IPMAN. To satisfy the need of traffic manager for the control of 'IP Service' , an innovational Service Provide Model is proposed .This article innovation spot has following points: It proposes the traffic control through analysis of the data flow characteristics instead of traditional QOS technology based on Diffserv and IntServ. This innovational Service Provide Model optimizes quality of network. The P2P data flow is efficiently controlled based on the analysis of P2P Application characteristics and dynamical adjustment of flow behavior instead of blocking or packet digest application detection .Referred to the self-seminal characteristic of IP network ,we make macroscopical management with fuzzy congestion control of IPMAN.Flow-state Aware Network Detection and Traffic Control Based on IPMAN is deployed in practical project ,which is one of the earliest successful cases in China. It fills the present IPMAN Traffic Engineering domain blank for China Telecom and has the reference meaning in practical project .It contributes the strategic factor of building the IPMAN Service Level System for China Telecom.
Keywords/Search Tags:QOS, Diffserv, P2P, BT(BitTorrent), Flow State Aware, Self-Similar, Fuzzy Congestion Control, IPMAN
PDF Full Text Request
Related items