MOVIES: indexing moving objects by shooting index images

Abstract

With the exponential growth of moving objects data to the Gigabyte range, it has become critical to develop effective techniques for indexing, updating, and querying these massive data sets. To meet the high update rate as well as low query response time requirements of moving object applications, this paper takes a novel approach in moving object indexing. In our approach, we do not require a sophisticated index structure that needs to be adjusted for each incoming update. Rather, we construct conceptually simple short-lived index images that we only keep for a very short period of time (sub-seconds) in main memory. As a consequence, the resulting technique MOVIES supports at the same time high query rates and high update rates, trading this property for query result staleness. Moreover, MOVIES is the first main memory method supporting time-parameterized predictive queries. To support this feature, we present two algorithms: non-predictive MOVIES and predictive MOVIES. We obtain the surprising result that a predictive indexing approachconsidered state-of-the-art in an external-memory scenario-does not scale well in a main memory environment. In fact, our results show that MOVIES outperforms stateof-the-art moving object indexes such as a main-memory adapted B x -tree by orders of magnitude w.r.t. update rates and query rates. In our experimental evaluation, we index the complete road network of Germany consisting of 40,000,000 road segments and 38,000,000 nodes. We scale our workload up to 100,000,000 moving objects, 58,000,000 updates per second and 10,000 queries per second, a scenario at a scale unmatched by any previous work.

An experimental analysis of iterated spatial joins in main memory

“…Thus we need not support updates and can use efficient bulk-loading methods. We evaluate four different choices of static index: the R-Tree [14,22], the CR-Tree [21], the Linearized KDTrie [9,29] and a Simple Static Grid [38]. R-Tree.…”

“…On one hand, we can use methods specifically designed for moving objects, such as the TPRTree [37]. On the other hand, if queries and updates can be processed in bulk, then it may be faster to re-execute a static join or rebuild a static index to deal with every batch of queries and updates [9,40]. It is clear that rebuild approaches should outperform incremental updates when batch sizes approach the total number of objects.…”

“…Furthermore, several recent studies have argued that many movingobject applications, including more traditional location based services such as flight tracking, can tolerate some query staleness [9,32]. However, these applications require very low response times even in the presence of significant query and update rates.…”

“…First, there have been a tremendous number of join algorithms proposed in the literature, many of which have never been compared directly. A number of existing algorithms are based on an index nested loops join, and thus require that we select from a wide variety of spatial indices for the inner relation [9,14,21,22,29,30,37,38]. Other approaches process the entire join in bulk, relying on lightweight auxiliary structures instead of a full-blown spatial index [3,5,17,31].…”

Parallelizing Data-Centric Programs

“…We have observed that all of the example applications above, as well as many others, can be run completely in main memory. Furthermore, several recent studies have argued that many moving-object applications, including traditional location based services such as flight tracking, can tolerate some query staleness [DBVS11,ŠRJŠ11]. However, these applications require very low response times even in the presence of significant query and update rates.…”