Combined Cache Policy For Web Service Workflow Execution Acceleration

In recent years, to compose a set of web services as a service process is becoming a common practice. When a service process is executing, a service process engine is responsible for service invocations in terms of the process logics. In order to accelerate the executions of service processes, the invocation results for services can be cached to facilitate future reuse so that the re-invocations of the services can be possibly avoided.A combined cache policy based on the service relationship modeling is proposed. The cache problem of service process is formalized firstly. The data structure for this policy based on chained table is then defined. How to refresh the cache is also discussed.At first, we introduced the service processes cache technology, emphasized on the scientific workflow intermediate data storage as well as semantic caching technology. Then given cache partitioning algorithm based on the consideration of time. The algorithm cache partitioning multiple was used to achieve cached combination. After that, we showed the detailed of the cache block, introduced their structure and their match strategy, maintenance strategy and cost and benefit. At last, we introduced a service process instance watching movies. The simulation is made to get the service process execution time update cost, including update time and cache size and network overhead(the cache visit times) after using the combined cache policy.Experiments’results show that the execution speed is increased comparing with the policy of independent service cache. At the same time, cache visit times is decreased.