CGVL:An Hierarchical Locking Mechanism

Abstract

Abstract:The Main-memory database systems are emerging as the new backbone for carrying out the  business processes. Besides flat relational data representations the hierarchical data structure  are essential for modern applications.We devise a new locking and versioning approach for hierarchies that is deeply integrated into the database.The existing Multiple granularity Locking Protocol are either too coarse or too fine grained when multiple threads request for the data item in the various part of the hierarchy.Thereby, leading to unpredictable, inefficient amount of concurrency.We therefore propose an efficient locking mechanism namely Collaboartive Granular Version Locking (CGVL) which follows an integration approach by combining the  flexibility and query capabilities  features from the Multiple Granularity locking (MGL) which is being successfully combined with the high read concurrency offered by Multi Version.We have been able to improve the concurrency by maintaining a series of version at each granular level so as to allow the simultaneous execution of read and write operation on a data item. Although many Multi Granularity algorithm have been proposed in literature that improve the traversal cost,but to the best of our knowledge none of them focus on avoiding the waiting state by read-only transactions.