| Summary: | 碩士 === 國立中興大學 === 資訊科學與工程學系 === 103 === With the development of the network, data transmission has become a main issue when it comes to network traffic. In order to improve the quality of service, many websites use the Content Delivery Network (CDN) technology to accelerate data access and prevent surrogate servers from overloading. In CDN scenarios, a good request routing algorithm selects the best surrogate server for clients’ request, and it takes two factors to select a server－the network and surrogate servers. Typical metrics are network latency, hop count, available Internet bandwidth, and surrogate server load. Selecting surrogate servers efficiently in CDN has always been a challenge. In the past, load balance, Round Robin and Network distance Routing were proposed selecting surrogate servers in Content Delivery Network. In this thesis, a Decision-making mechanism for improving load balance (DMM-ILB) is adopted to solve the problem. Our idea is to exclude the surrogate servers, which the round trip time is higher than average value, also, increase the number of filtered surrogate servers during peak hours at the same time. To re-route requests a suitable surrogate server. DMM-ILB provides a stably high throughput with low error rate while delivering data to clients. The simulation results show that DMM-ILB method can achieve a better performance than other algorithms.
|
|---|
