Replicated Web Services: A Comparative Analysis of Client-Based Content Delivery Policies
Marco Conti, Enrico Gregori, Willy Lapenna · Lecture notes in computer science · 2002
The increase of the Internet users and web applications leads to the need for more reliable and faster Web services. Different techniques were developed to address this issue and to provide a better QoS for the Internet users. Among them, an important role is played by the replication of Web services. Replication of Web services is obtained by implementing a Web service with several Web servers. This replication can be deployed either locally (e.g. by a cluster of computers) or geographically (e.g. servers distributed sparsely in the Internet). In this paper we deal with geographical replication. Two main approaches are currently used for geographical replication: client-side and server-side approach. We focused our analysis on the client-side approach. We classified and contrasted, qualitatively and quantitatively (via simulation), different client side techniques to find the pro and cons of each approach with the aim to identify the best solutions for content-delivery systems. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.