Framework for managing distinct versions of data in relational databases

Abstract

Despite the growing number of different database and storage technologies, relational databases still support significant percent of internet and intranet applications due to their effectiveness in processing complex queries and capabilities to provide data integrity in concurrent data access.Most of modern relational DBMS develop their concurrent access models to meet the requirements of transaction consistency, but at the same time they lack tools that provide usability in complex scenarios of concurrent data processing, that require storing and retrieving multiple versions of the same tuple.This paper describes the high-level framework to manage multiple co-existing versions of data tuples, and to support complex workflows of concurrent data modifications, including “branched” versioning.The different aspects of design and implementation of such a system are discussed.