Abstract:

In the era of mobile big data, traditional location based service (LBS) techniques face new challenges such as lack of system scalability and performance. We first propose a query framework based on Storm according to the characteristics of LBS applications. Then, we design continuous parallel range query algorithms based on Storm to optimize query performance. As for the consistency problem in the distributed environment, we design a distributed lock service based on ZooKeeper to guarantee the correctness of query results. Furthermore, we propose a cacheoptimized algorithm based on TimeCacheMap and two caching strategies for the timeconsuming problem of accessing database in the parallel continuous range query algorithm based on Storm, so as to reduce the overhead of accessing database and improve query efficiency.

Key words: Twitter Storm, continuous range query, parallel query processing, optimization technique