Concurrent updating transactions on versioned data

Abstract

Modern database applications increasingly often require access to historical versions of the database. Storing such multiversion data in a single-version B + -tree database index is inefficient, especially for key-range queries. In this article, we present an index structure called the concurrent multiversion B + -tree (CMVBT) for efficiently storing and querying multiversion data.The CMVBT structure uses an asymptotically optimal transactional multiversion B + -tree (TMVBT) index as the main data storage, and a separate B + -tree index called the versioned B + -tree (VBT) to hold the updates of active transactions. The updates of committed transactions are moved, one transaction at a time, from the VBT into the TMVBT. This organization of two separate index structures allows us to maintain the asymptotic optimality guarantees of the TMVBT even in the presence of concurrent updating transactions.We provide concurrent algorithms for updating and reading the CMVBT structure. Our CMVBT algorithms can be used with the standard snapshot isolation concurrencycontrol and ARIES-based recovery algorithms to allow multiple read-only and updating transactions to operate concurrently on the structure. Transaction rollback is also supported for all updating transactions, either entirely or up to a preset savepoint.

A two-part multi-algorithm concurrency control optimization strategy for distributed database systems

“…All data items are important parts of the DDBS, but performance-wise, Our algorithm suggests a categorical classification of data items in terms of serializability requirements into crucial data items that should strictly follow the serializability scheduling mechanism, examples of such data include banks customers' credit, money, and balance transactions, and into those that can follow a more relaxed form of serializability for employee information in the form of a snapshot isolation mechanism such as the mechanism used by Haapasalo et al (2009) with a concurrent multiversion B+-tree to allow multiple concurrent readonly transactions to access historical states of the database, or even updating transactions using relaxed forms of serializability scheduling while maintaining security in both categories. In each tier, hence, the concurrency control algorithm has a master site to minimize cost in terms of processing capacity and synchronization messages as much as possible by increasing the autonomy of concurrency control algorithms or the mechanism that oversees its handling.…”

Mv-Idx

“…under MVAS every index record has a va- lidity period marked by a start and end timestamp. In a series of papers Haapasalo et al describe an approach called Multi-Version BTree [8,5,6,7] aiming to provide a general index structure for MV-DBMS. In [3] we describe the Snapshot Isolation Append Storage multi-version index which identifies tuple versions of a single data item using a single virtual ID.…”