Multiple Granularity Locks for the KBMS Environment

As ever-larger knowledge bases (KBs) are being built, knowledge sharing becomes an aspect of paramount importance in Knowledge Base Management Systems (KBMSs). In this paper, we propose a way of controlling knowledge sharing by means of the LARS (Locks using Abstraction Relationships’ Semantics) approach for concurrency control in KBs. LARS synchronizes transactions through many different granules of locking, which are based on the semantics of the abstraction relationships commonly used in knowledge representation approaches. LARS supports a higher degree of potential concurrency in that it maintains different logical partitions of a KB graph, a means for representing KBs, and offers many lock types to be used on the basis of each one of the partitions. By such a way, LARS captures more of the semantics contained in a KB, through an interpretation of the (abstraction) relationships between objects, profits from such semantics for synchronizing the transactions, and thus makes feasible the exploitation of the inherent parallelism in a knowledge representation approach.