Parallel pointer-based join techniques for object-oriented databases
Daniel F. Lieuwen, David J. DeWitt, M. Mehta · 2002
Four parallel pointer-based join algorithms for set-valued attributes are described. Such joins will be common in next-generation object-oriented database systems. It is shown that while algorithms based on hybrid-hash provide good performance, algorithms that require less replication will often produce as good or better performance, especially if each set-valued attribute references a small number of nodes.>