OMCAT: Optimal Maintenance of Continuous Queries' Answers on Trajectories

OMCAT is an extension of our previous work on Context-Aware Maintenance of Continuous Queries(CAT), where CAT was a proof of concept that triggers available in the commercial ORDBMS can be used to maintain spatio-temporal range queries in MOD. However, in OMCAT, we show the impact of different trigger semantic/context dimensions on the efficient reevluation of continuous queries. We also show how the query reevluation can be optimized by means of intra-query level heuristics that utilize the spatio-temporal context information embeded in the moving object trajectories, as well as the various types of pending queries.

Introduction:

In Location-Based Services, it is desirable to maintain corrent answers to a set of pending continuous queries efficiently. The functionality of query maintenance is provided by the so-called Moving Object Database(MOD). In our settings, MOD stores the trajectories that describe the motion of moving objects, which are pertaining to the future. It also stores the various types of pending queries on the moving objects. The queries considered include: static range query, dynamic within distance query and dynamic k-nearest neighbor query. The database also stores traffic abnormality that may affect the planed motion of one or many moving objects, which may result in a call for certain query reevaluation. These traffic abnormalities may be reported on-line by various sources and have an immediate impact on the pending query maintenance. When the database receives any traffic abnormality, our programmed triggers tht monitor the pending query will fire to perform any necessary query reevaluation. For a more serious treatment on our study of continuous query maintenance for trajectories, please refer to [1], [2]. OMCAT demonstrates the process of query maintenance, from the moment a traffic abnormality is reported to the system, to the time when all query answers are brougt up-to-date again. 4000 trajectories are pre-stored in the database, representing 4000 moving objects in the urban area of city of Chicago. 9 continuous queries are also pre-stored in the database to present the best demonstration effect. However, the user is free to specify any query of its own. The user can also specify a region of traffic abnormality with certain delay effect. After OMCAT receives the user input, it sends this input to the back-end database and performs query (re)evaluation. OMCAT then receives the result from the database and update its presentation to the user.

DEMO screenshot:

Figure 1 illustrate the architecture of OMCAT.

Figure 2 shows the main GUI interface of OMCAT.

Figure 3 is the screenshot after a user specifies certain query on the moving objects.

Figure 4 shows the result after a user-specified traffic abnormality is submitted to the database. The query answers are updated. Note that the user can select the manner of query reevaluation, using either the naive way or our optimized approach.

Instructions for downloading and running OMCAT

Our continuous query maintenance prototype is built on top of Oracle Spatial, although it can be implemented for any commercial database that supports User-Defined datatypes and User-Defined functions, and use triggers or similar rule management system to provide reactive behavior. The query maintenance logic and programmed trigger are all implemented in PL/SQL and the source code can be downloaded here.

To run our query maintenance system yourself, an Oracle Database of version 9i or higher is required. To run OMCAT and see the visual effects, you will also need to have Oracle Application Server Mapviewer installed. Currently due to the limited computing resource, we do not provide an on-line DEMO although one is possible. Interested visitor can contact the author at: hdi117-AT-ece.northwestern.edu for further assistance.

References:

1. Context-Aware Optimization of Continuous Range Queries Maintenance for Trajectories, Goce Trajcevski, Hui Ding and Peter Scheuermann, in Proceeding of MobiDE'05.
2. Context-Aware Optimization of Continuous Query Maintenance for Trajectories, Hui Ding, Master's thesis, Northwestern University, 2005.
Your are the 's visitor of OMCAT.