OneService - Generic Cache Aggregator Framework for Service Dependent Cloud Applications

Current big data cloud systems often use different data migration strategies from providers to customers. This often results in increased bandwidth usage and herewith a decrease of the performance. To enhance the performance often caching mechanisms are adopted. However, the implementations of these caching mechanisms are often dedicated solutions for specific applications and/or use case scenarios. The adoption of different caching implementations within the same system leads to different problems including increased maintenance overhead, decrease of reuse, reduced adaptability, and resource allocation problems. To overcome these problems, in this paper we propose the so-called OneService Framework, which provides a generic cache aggregator mechanism that can be used with different cache storages to fetch and distribute the data from various providers. The framework as such helps to increase reuse, support adaptability, resolve the resource allocation problems, and enhance the overall performance of the system. We discuss the overall design of the framework together with the basic implementation concerns. The framework is illustrated for analyzing the maintenance, reusability and cost of MSN backend services.