Data Loading and Caching Strategies in Service-Oriented Enterprise Applications

The use of distributed functionality through service interfaces provides enhanced application functionality. In contrast to traditional desktop applications, enterprise applications have to consider performance aspects as a result of the distributed application setup. Cost models identified data transmissions as cost-drivers in service-oriented applications because of their substantial size. An object cache helps to avoid redundant service calls and it improves perceived application performance. We present implementation details about our abstract application layer for consumption of enterprise services providing dynamic load and transparent caching methods. Dynamic loading of business object data facilitates comfortable use of enterprise services by clients without expertise on service implementations. We examine various server- and client-side caching strategies and implemented a performance testbed. Our performance results are discussed and solutions for enterprise architectures are drawn.