Relaxed AVL trees, main-memory databases and concurrency

We consider the use of search trees to represent the dictionary aspects of a main-memory database in a concurrent environment. Efficiency considerations require that the trees be balanced and that operations on a search tree should not block too large a part of the tree for too long a time. These two requirements conflict, since rebalancing a tree after an update can, in a straightforward implementation, block too large a part of the tree. We would prefer that au search operations reserve only a small, fixed-size part of a tree at all times. We propose a new, elegant solution for this problem based on the notion of relaxed AVL trees and the decoupling of updates and rebalancing. The main advantage of our solution is that the implementation of the dictionary operations in a concurrent environment is as simple as their implementation in a sequential environment, whereas previous concurrent solutions are more descriptively complex.