Hash-based and index-based join algorithms for cube and ring connected multicomputers

Parallel hash-based join algorithms for a shared-everything environment

“…One environment where join algorithms have been studied is machines with large main memories 6,19]. Another popular environment for this work has been distributed-memory, or shared-nothing multiprocessor systems 1,3,7,9,12,16,20]. In most of these studies it has been found that, for the join of large sequential les, hash-based algorithms outperform the nested loops and the sort-merge algorithms.…”

Parallel join for IBGF partitioned relational databases

“…They are expected to perform equally well. Given a multi-unit disk storage system, the issue of how to map partitions in order to maximise the concurrency, during a join, is closely related to the efficiency [12]. The partition mapping is controlled by the size and topology of the storage subsystem, the parallel computing system, as well as the relation characteristics.…”

Multiway Equijoin Query Acceleration Using Hit-Lists

“…Building on this and the work of others such as Kitsuregawa (1983), DeWitt et al (1984), and Shapiro (1986) developed non-persistent, hashed bit filters in software, and brought hash joining to its current state of development, hybrid hash. Valduriez and Gardarin (1984), DeWitt and Gerber (1985), and Omiecinski and Lin (1989) also extended these ideas to the partitioning of multiprocessor joins. Recent research by Zeller and Gray (1990) has involved the implementation of hashing functions in practice.…”