Multi-dimensional dynamic bucket index based on mobile agent system architecture

Abstract

This paper describes the idea of a bucket index employed for processing of intensive reading streams coming from huge telemetry networks. This data structure answers approximate spatio-temporal range queries concerning utility usage in user selected region and time, The index structure continuously adjusts to data distribution changes and, as opposed to traditional indexing methods, is capable of processing of the updates on the fly. A stochastic prediction model is also used to estimate utility usage in the near future. The presented indexing technique is implemented in a distributed system based on mobile agents. The mobile architecture is used to control the workload of network hosts. © Springer-Verlag Berlin Heidelberg 2006.