Optimized Caching Of Service Processes

With the development of technologies like Internet and cloud computing,as well as the advent of the Web 2.0 era,Internet is gradually transforming from an information-sharing platform to a service-sharing platform,where Web service is the main vehicle,via which a growing number of scientific institutions and Internet companies share their data and services.Since the limited function of a single Web service,composing a set of Web services as a service process is becoming a common practice.In that Web service processes involve multiple service invocations,it incurs a huge network transport cost,which could become a bottleneck in the execution.In order to solve this bottleneck,there are already some studies and technologies in the fields of process caching and Web service semantic caching,but there are still several issues:(1)The operation and parameter's type of the Web service request are not considered in Web service semantic caching;(2)The expiration of the Web service responce is not taken into consideration in the process caching like Mashup caching;(3)The structure of service processes is not taken full advantage of in process caching.For these problems,this paper takes the expiration of service results into account,models the cache structure of processes and the query of Web services,and caches serveral structurally related services as a whole to reduce the cache space consumption as well as to improve the cache performance.In addition,the paper studies the space assignment and cache replacement of processes caching as well as optimizations for caching multiple processes.The main work of this paper is reflected in the following aspects:1.A model supporting service process caching is proposed,where formal definitions of service processes and related elements,and the calculation of related metrics are given,followed by a model of generic Web servie queries and the application of semantic caching in Web services.2.A algorithm based on cache blocks to find the best caching schema and a local quering global replacing strategy taking into consideration the caching profit,size of the cache record and size of the cache block are developed for caching a single service process.Experiments of comparison based on a real service process to verify the correctness and effectiveness of the algorithm and strategy are also performed.3.For cahing multiple processes,the sharing of common cache blocks are studied.And a replacement strategy based on the strategy of caching a single process which also considers the dynamic changes of the execution frequency of different processes is proposed.Experiments of comparison to verify the effect of sharing common cache blocks and the strategy are performed.