Christos A. Polyzois

Management of a remote backup copy for disaster recovery

A remote backup database system tracks the state of a primary system, taking over transaction processing when disaster hits the primary site. The primary and backup sites are physically isolated so that failures at one site are unlikely to propogate to the other. For correctness, the execution schedule at the backup must be equivalent to that at the primary. When the primary and backup sites contain a single processor, it is easy to achieve this property. However, this is harder to do when each site contains multiple processors and sites are connected via multiple communication lines. We present an efficient transaction processing mechanism for multiprocessor systems that guarantees this and other important properties. We also present a database initialization algorithm that copies the database to a backup site while transactions are being processed.

Issues in disaster recovery

The increased demand in recent years for continuous operation of databases has created interest in remote backups, which allow systems to continue their operation, even in the presence of disasters. The motivation for remote backups and their applicability to various systems are discussed. A remote backup can ensure continuous operation, even in the presence of extensive failures that may render an entire site inoperational and for which local replication may be inadequate. In particular, the geographic separation of two copies helps isolate failures that might otherwise affect the backup copy, as well as the primary. Some of the issues involved in the design, implementation, and evaluation of remote backup mechanisms are presented. Consistency is a major interest for remote backups: a recovery mechanism may produce either a consistent backup copy or an inconsistent one. Consistency will be highly desirable in most cases, since application programs are usually written under the assumption that they will operate on correct data, and error handling in these programs is rarely comprehensive. Some criteria that can be used for evaluating and comparing various remote backup mechanisms are presented. Topics discussed are taxonomy, implementation choices, and comparison metrics.<<ETX>>

Evaluation of remote backup algorithms for transaction processing systems

A remote backup is a copy of a primary database maintained at a geographically separate location and is used to increase data availability. Remote backup systems are typically log-based and can be classified into 2-safe and 1-safe, depending on whether transactions commit at both sites simultaneously or they first commit at the primary and are later propagated to the backup. We have built an experimental database system on which we evaluated the performance of the epoch algorithm, a 1-safe algorithm we have developed, and compared it with the 2-safe approach under various conditions. We also report on the use of multiple log streams to propagate information from the primary to the backup.

Overview of disaster recovery for transaction processing systems

An overview is given of the major issues involved in maintaining an up-to-date backup copy of a database, kept at a remote site. A method is presented for performing this task without impairing the performance at the primary site. The method is scalable, and it is particularly suitable for multiprocessor systems. The mechanism is relatively straightforward and can be implemented using well-known concepts and techniques, such as locking and logging.<<ETX>>

Processing of read-only queries at a remote backup

Remote backup systems are often used to provide high data availability. Updates are typically propagated to the backup via a log, which decouples the backup from the primary. We show that this decoupling can lead to efficient installation of updates in batches and efficient processing of read-only queries, by eliminating or reducing access conflicts between updates and queries. We present several methods for query processing at the backup and evaluate their performance analytically.<<ETX>>

Restart services for highly available systems

This paper proposes a design methodology for building highly available systems. In addition, we describe a set of operating system services that can be used to achieve this goal. The techniques described are intended for a parallel environment and can be generalized for any distributed system. We describe a methodology for providing basic services for high availability, specific services for restart and an implementation of these services.