Replication based on objects iteration frequency and load using a genetic algorithm under a content distribution network

One of the important problems related to content management in content distribution networks (CDNs) is an optimal placement of outsourced content in CDN´s servers. Taking into account that this problem is NP complete, a heuristic method should be developed. All the approaches developed so far take the iteration frequency or the workload as criteria. In this paper, a novel, evolutionary and non-parameterized technique has been proposed which uses the genetic algorithm to place content in CDN´s servers. In this technique, objects are placed in servers according to iteration frequency, load and delay generated by the objects. Through a detailed simulation environment, we show that the proposed technique improves significantly the mean response time and hit rate while keeping the CDNs´ servers´ load at a very low level.