Hypercube-based caching: a multidimensional semantic architecture for efficient information mediation

Using semantic caching to meet user requests has been of great interest. Many information systems, including distributed databases, data warehouses and mediation systems encounter challenges related to network unavailability and congestion. In such cases, semantic caches can be extremely useful. In a semantic cache, queries that have already been answered are stored with their respective formulas to be used in future demands. In this paper, we propose a new semantic query caching for mediation systems. In our caching model, cached regions are semantically represented on multiple semantic dimensions in the form of hyper-cubes. We provide an algorithm to project queries into their own hyper-cubes and store them in regions. We also examine when it is advisable to use a given hypercube in the cache to partially answer a user query and when regions should be skipped to reduce response times and obtain relevant answers.