Besides a community of open-source developers, more than a dozen companies chipped in on development for Version 8.0, including Red Hat Inc., Fujitsu, Afilias Ltd., Software Research Associates Inc., 2nd Quadrant Ltd. and Command Prompt Inc.
The support for Windows, a feature customers have been requesting for many years, is widely seen as the most important feature of the new release. "This will really take PostgreSQL to the next level, because Windows seems to be a major deployment platform for open-source [deployment], and theyve been missing out" on that, said Noel Yuhanna, an analyst at Forrester Research Inc., in Santa Clara, Calif.
Native Windows support means that PostgreSQL aficionados will no longer need to run the database with an emulation layer, which will in turn dramatically improve performance.
Indeed, it could well be the feature that propels the database out of the realm of developers, who have long embraced it, and into the realm of corporate IT management, Yuhanna said. "I think its popular with the developers, not as much with the mainstream IT organizations," he said. "Commercial DBMSes [database management systems] still hold the key data for most organizations. And the adoption really comes from the developers. Once the developers and administrators start liking the product, start using it and understanding it, then comes the next level, of using it for small, mission-critical deployments, and thats where customers see the advantages" of open-source and of PostgreSQL in particular, he said.
New features include Savepoints, an SQL-standard feature that allows parts of database transactions to be rolled back instead of aborting entire operations. Point inTime Recovery is another new feature that allows for full data restoration from automatic and continuously archived transaction logs—an alternative to hourly or daily backups.
Also new to Version 8 are Tablespaces, which allow for placement of large tables and indexes on individual disks or arrays—a feature that improves query performance and is particularly relevant with large data warehousing systems.
Improved memory and I/O have also resulted from the Adaptive Replacement Cache algorithm, a new background writer, and a new vacuum delay feature.