Font Size: a A A

A Study Of Cooperative Optimization On Traffic Engineering And Server Selection

Posted on:2011-11-27Degree:MasterType:Thesis
Country:ChinaCandidate:Y WenFull Text:PDF
GTID:2178330338489733Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
P2P streaming media providers are emerging rapidly. The flows increase but the profits shrinke for ISP greatly because the disorder of P2P streaming media flow has used the network traffic greatly, occupyed the bandwidth of importance, and declined the network performance. According to a series of problems caused by the P2P streaming media services, we must optimize P2P streaming media and strengthen the cooperation of ISP and content provider.ISP is engaged in solving traffic engineering, and CP is mainly to solve server selection. The relations of of traffic engineering and service selection include independent optimization, partial cooperation, and full cooperation. We solve the three mathematical models of the three relationships with the Lagrange dual decomposition method in this thesis. Constraint condition is relaxed using Lagrange oprator, and the original problem becomes a main problem and several statures, and variables are updated using gradient method. In consideration of the actual algorithm convergence speed and computational complexity and other factors, another solution is obtained with direct derivation and mapping method. The ability of links and servers is ignored, and the variable is derivated. The result is projected to [0,1], and then the network is updated until it's converged.To verify the correctness of the algorithm and test the degree of performance improved, a program design is completed in Planetsim simulation platform, including the initialization, optimization and outputting modules. In the initialization module, the servers and storage contents are configured; the users and the files are choosed, the state of the links and paths are configured. In the optimization module, it is calculated in accordance with the algorithm, and then the variables and networks are renewed until the network is stable. In the output module, exports are designed, and the values of the convergence and parameters are written to a specified file. Finally, the algorithm is tested in actual networks with different topologies and the results are analyzed. The superiority of the cooperative model and the correctness of solution to the algorithm are verified.
Keywords/Search Tags:P2P streaming media, traffic engineering, server selection, full cooperation, convex optimization
PDF Full Text Request
Related items