Latency-Optimized Traffic Provisioning Algorithm In CDNs

With the explosion of network data transmission,Content Delivery Networks(CDNs)cache and deliver hundreds of trillions of user requests each day from thousands of servers worldwide.The traffic served by CDNs can be partitioned into thousands of domain requests,each with different user access patterns,popularity distributions,performance requirements,etc.The goal of CDN is to host and deliver content and applications to clients around the world with high availability,performance,and scalability.One critical problem is how to reduce the time between request and response for content as perceived by the end-user since optimizing the request-response time of CDNs can significantly improve the quality of experience(QoE)of a variety of applications built upon it.This work will explore the traffic provisioning problem in CDN,and our goal is to minimize the end-user latency by designing a sensible traffic provisioning algorithm.The main work of this thesis is summarized as follows:●A model is developed for latency-aware traffic provisioning that assigns traffic classes to servers to minimize the end-user latency.●An algorithm is proposed that can accurately calculate the hit rate in any specified cache space after the traffic of each domain is mixed in any proportion in real production data.●This paper proposes a domain name traffic distribution algorithm for CDN servers that comprehensively considers traffic load balancing,server cache hit rate,and delay factors.●The algorithm is applied to metro-level traffic provisioning where the domain traffic assigned to servers within a cluster is re-provisioned to minimize end-user latency.The results show that LAMM has a 10%improvement over the most widely used consist hashing based schemes and consistently performs well under various settings and environments.