Framework for Horizontal Scaling of Map Matching: Using Map-Reduce

Abstract

Map Matching is a well-established problem which deals with mapping raw time stamped location traces to edges of road network graph. Location data traces may be from devices like GPS, Mobile Signals etc. It has applicability in mining travel patterns, route prediction, vehicle turn prediction and resource prediction in grid computing etc. Existing map matching algorithms are designed to run on vertical scalable frameworks (enhancing CPU, Disk storage, Network Resources etc.). Vertical scaling has known limitations and implementation difficulties. In this paper we present a framework for horizontal scaling of map-matching algorithm, which overcomes limitations of vertical scaling. This framework uses Hbase for data storage and map-reduce computation framework. Both of these technologies belong to big data technology stack. Proposed framework is evaluated by running ST-matching based map matching algorithm.