Parallel evaluation of the transitive closure of a database relation

Abstract

Parallelism is a promising approach to high performance data management. In a highly parallel data server with declustered data placement, an important issue is to exploit parallelism in processing complex queries such as recursive queries. In this paper, we consider the transitive closure of a database relation as a paradigm to study parallel recursive query processing. And we propose two new parallel algorithms for evaluating the transitive closure of a relation in a parallel data server. Performance comparisons based on an analytical model indicate the superior response time of the parallel algorithms over their centralized version. With one hundred nodes, performance gain is between one and two orders of magnitude. One parallel algorithm provides superior response time while the other exhibits better response time/total time trade-off.

Parallel processing of graph reachability in databases

“…By curve-fitting these figures into our broadcastnetwork formulas we obtained that in our environment C1, C2, and C3 are approximately 0.02, 0.0032, and 0.001, respectively. In other words, for these C's our formulas 5,19,20, and 21 predict speedups that are close to the ones that we actually obtained. For example, for these C's, our for- mulas predict a speedup of 1.652893 for SSR1 on 2 processors (in contrast to 1.75 obtained experimentally), and a speedup of 2.904877 for SSR2 on 3 processors (in contrast to 2.68 obtained experimentally).…”

“…Similar strategies have been used in Cheiney and Maindreville,~3) and Valduriez and Koshafian, (5) for parallel computation of the transitive closure. The strategies divide the computation work by partitioning the relation representing the input graph, and assigning to each processor the responsibility for producing only a fraction of the reachable nodes.…”

“…(see Refs. [2][3][4][5]. However, the topology of the input graph has proved to be problematic, since there is no model in which the speedup can be represented as a function of the topology.…”

Main memory evaluation of recursive queries on multicore machines

“…The novel performance findings presented in this paper are hardly surprising in view of the fact that the most previous studies date back to the late 80's and early 90's [6]- [9], and there has been much progress in multicore systems since then. Moreover, we assume here that all our data resides in main memory, whereas past studies on recursive query evaluation [4]- [17] often assumed a database-oriented environment with data residing on secondary storage, whereby query evaluation algorithms were designed to reduce I/O costs rather than inmemory evaluation costs.…”

Algebraic optimization of recursive queries

“…A body of research has been performed on this subject [3,7,23,28], and transitive closure is being supported by some database systems [37]. Recently, one has started to study the use of distributed computation for transitive closure queries [19,40,44]. This is, however, still a difficult topic, and it seems that one should have some knowledge about the application domain to really benefit from parallel processing, as in [23,24,25].…”