A locking scheme for associative retrieval

Abstract

The general problem of concurrency control for database systems has been studied intensively. However, the studies often ignored the underlying file systems. A locking scheme for multi-key hashed file structures is presented in this paper. First, we define a transaction model for such file structures, then we present techniques for conflict detection as well as a new locking scheme. The locking scheme combines aspects of both physical and predicate locking schemes. It places locks on buckets while guaranteeing record-level access; this results in increased concurrency with minimal overhead.