Implementing and Evaluating R-Tree Techniques on Concurrency Control and Recovery with Modifications on Nonspatial Domains

Upon review of the present applications that work on use of database for spatial data, it is identified that the same needs are to be incorporated in the database management systems for better support on these products. This research talks about one of those techniques in the context of handling spatial data incorporating its nonspatial element. Spatial objects are mostly handled via a minimum bounding box in most popular spatial access methods. This kind of generalization and approximation is fast but inaccurate for answers to queries. Many researchers have already worked on finding better minimum geometrical shape for a spatial object. This research takes it further and implements one such better method minimum binding circle (MBC). Apart from R-link tree, no other research has been done in incorporating a nonspatial element into the spatial object. In R-link too it is at a very fundamental level wherein the element inserted is a logical sequence number to be used for sequencing the nodes in the tree and has no relevance in the database. However, this research takes it further and introduces NS link tree (nonspatial) with minimum bounding circle and adding relevant nonspatial data at each point to reduce the number of query results to the database, thereby proving that the links to the database based on the queries are considerably reduced. Concurrency control is maintained through a priority queue. Separate log files are used to handle recovery.