Abstract: In this paper, we show that spatial joins are very suitable to be processed on a parallel hardware platform. The parallel system is equipped with a so-called shared virtual memory which is well-suited for the design and implementation of parallel spatial join algorithms. We start with an algorithm that consists of three phases: task creation, task assignment and parallel task execu-tion. In order to reduce CPU- and I/O-cost, the three phases are processed in a fashion that pre-serves spatial locality. Dynamic load balancing is achieved by splitting tasks into smaller ones and reassigning some of the smaller tasks to idle processors. In an experimental performance compar-ison, we identify the advantages and disadvantages of several variants of our algorithm. The most efficient one shows an almost optimal speed-up under the assumption that the number of disks is sufficiently large. Topics: spatial database systems, parallel database systems 1

Global clustering has rarely been investigated in the area of spatial dambase systems although dramatic performance improvements can be achieved by using suitable techniques. In this paper, we propose a simple approach to global clustering called cluster organization. We will demonstrate that this cluster organization leads to considerable performance improvements without any algorithmic ovedxad. Based on real geographic data, we perfm a detailed empirical performance evaluation and compare the clusterorganixation to other organization models not using global clustering. We will show that global clustering speeds up the processing of window queries as well as spatial joins without decreasing the performanceoftbeinsertionofnewobjectsaodofse, lective queries such as point queries. lhe spatial join is sped up by a factor of about 4, whereas non-selective window queries are accelerated by even higher speed up factors. 1

The basic concept for processing spatial joins consists of two steps: First, the spatial join is performed on the minimum bounding rectangles of the objects by using a spatial access method. This step provides a set of candidates which consists of answers (hits) and non-answers (false hits). In the second step, the exact geometry of the candidates is transferred from secondary storage into main memory and is tested against the join predicate. This step is called refinement step. It causes the main cost for computing a spatial join. In this paper, we introduce an additional filter step in order to reduce the cost of the refinement step. In this filter step more sophisticated approximations are used to identify hits as well as to filter out false hits from the set of candidates. For this purpose, we investigate various types of conservative and progressive approximations. The performance of the approximation approach is evaluated with data sets from real cartographic applications. The results show that this approach considerably reduces the total execution time of the spatial join.