Synchronizing shared abstract types
ACM Transactions on Computer Systems (TOCS)
Concurrency versus availability: atomicity mechanisms for replicated data
ACM Transactions on Computer Systems (TOCS)
Commutativity-Based Concurrency Control for Abstract Data Types
IEEE Transactions on Computers
Using semantic knowledge of transactions to increase concurrency
ACM Transactions on Database Systems (TODS)
Concurrency and availability as dual properties of replicated atomic data
Journal of the ACM (JACM)
Hybrid concurrency control for abstract data types
Journal of Computer and System Sciences
Using semantic knowledge for transaction processing in a distributed database
ACM Transactions on Database Systems (TODS)
Multilevel atomicity—a new correctness criterion for database concurrency control
ACM Transactions on Database Systems (TODS)
Concurrency control in a dynamic search structure
PODS '82 Proceedings of the 1st ACM SIGACT-SIGMOD symposium on Principles of database systems
Hi-index | 0.00 |
In a database system where the data is a collection of shared objects there can be concurrent access to the data by several transactions. Traditional concurrency-controls use a notion of conflict between pairs of operations (of which the transactions are composed) to ensure correctness. Herlihy has proposed the notion of a serial dependency relation over the set of operations to capture this idea of conflict. Since the smaller the conflict relation the more the concurrency, it is of interest to construct minimal serial dependency relations. In this paper, we give necessary and sufficient conditions for a pair of operations to be related by a minimal serial dependency relation. However, we go on to show that, in general, the problem of constructing a minimal relation is undecidable. We also show that some approaches advocated for constructing serial dependency relations are not feasible in general. We then provide a sufficiency condition for cases where the minimal relations are computable.