Based on extensive lab results, this white paper demonstrates that, in a highly-available disaster recovery environment using either HP StorageWorks Continuous Access EVA or Microsoft SQL Server 2005 database mirroring, checkpoint duration is critical to managing the solution behavior. Implementation of the best practices results in significant performance improvements:
- For Continuous Access EVA with heavy online transaction processing (OLTP) workloads—Modification of the checkpoint intervals results in a 14% improvement in overall transaction performance (measured in transactions per second [TPS]) and a 39% improvement in transaction response times.
- For SQL Server database mirroring—Modification of the checkpoint intervals results in a 59% improvement in overall transaction performance (measured in TPS) and a 52% improvement in transaction response times.
This solution provides the information needed to select the replication, mirroring, and/or high availability method(s) suitable for your environment; and the best practices for optimal configuration of the database, storage, and software. Intended for users who are running or plan to run disaster recovery solutions for SQL Server 2005 in an enterprise environment, the white paper provides the data needed to understand the options and limitations for implementing disaster recovery solutions on a clustered HP server and storage area network (SAN). Both array-based replication with HP StorageWorks Continuous Access EVA and host-based copying with SQL Server database mirroring are discussed, along with strategies for synchronous and asynchronous replication of SQL databases between primary and remote HP StorageWorks Enterprise Virtual Array (EVA) disk arrays.
By leveraging the recommendations and best practices, administrators can select the optimal replication technology for their environment and can accelerate implementation, thereby reducing cost and optimizing hardware and personnel resources.
|
Printable version
