TOF: A Throughput Oriented Framework for Spatial Queries Processing in Multi-core Environment

摘要

In this paper, we develop a Throughput Oriented Framework (TOF) for efficient processing of spatiotemporal queries in multi-core environment. Traditional approaches to spatial query processing were focused on reduction of query latency. In real world, most LBS applications emphasize throughput rather than query latency. TOF is designed to achieve maximum throughput. Instead of resorting to complex indexes, TOF chooses to execute a batch queries at each run, so it can maximize data locality and parallelism on multi-core platforms. Using TOF, we designed algorithms for processing range queries and kNN queries respectively. Experimental study shows that these algorithms outperform the existing approaches significantly in terms of throughput.