Content area

Abstract

The 2nd beta version of Microsoft Corp.'s SQL Server 7.0 is reviewed. Microsoft has implemented backward compatibility. The company has added an improved query processor, row-level locking, and expanding size limits. The query optimizer has been improved, and support has been added for distributed heterogeneous queries via ODBC and Microsoft's OLE DB interface. Support for row-level locking encompasses both data rows and index entries. This version of SQL Server works on everything from a notebook to a 16-processor symmetrical multiprocessor machine. Developers will like the improved Transact-SQL enhancements, the monitoring and debugging capabilities, and the integration with other Microsoft development tools, such as Visual C++.

Full text

Turn on search term navigation
 

CLEARLY OUTSCORING PREVIOUS versions, this beta release of SQL Server 7.0 shows that Microsoft is serious about playing pro ball with database rivals Oracle, IBM, and Informix. This version demonstrates good stability along with a nearly complete feature set. Current SQL Server 6.x customers and Windows NT shops seeking a Windows-based solution for decision support, data warehousing, or OLAP applications should keep an eye out for the production version when it arrives later this year.

View Image -

Microsoft still needs to tweak the performance a bit, add the final few features, and do the usual buffing and polishing prior to shipping. Aside from these foibles, SQL Server 7.0 exhibits impressive improvements in architecture, ease of use and manageability, and development options. But despite this achievement, it is too early to compare this beta version with the production releases of database rivals.

For sites that already are using SQL Server 6.x, upgrading to 7.0 is unlikely to cause many headaches, as Microsoft wisely has implemented backward compatibility. During my tests with the 6.5 and 7.0 versions I experienced no interoperability problems.

The architectural improvements in this release are particularly significant. The company has added a greatly improved query processor, row-level locking, and expanded size limits.

The query processor now supports parallel execution, index intersection, and hash joins, which are useful performance boosters when querying large data stores. The query optimizer also has been improved, and support has been added for distributed heterogeneous queries via ODBC and Microsoft's OLE DB interface.

Support for row-level locking encompasses both data rows and index entries. This is a particularly important enhancement because some enterprise-level applications, such as SAP's R/3, require row-level locking. The lock manager also seems to work better in this version and it dynamically adjusts itself depending on the environment.

The expansion of size limitations also is a welcome change. Although still fixed, database pages have increased from 2KB to 8KB, and a row can now expand to a tad more than 8,000 bytes. Tables can now have a maximum of 1,024 columns, as opposed to the previous limit of 250 columns. This will make SQL Server a more credible contender in large enterprise settings.

I also was impressed with the fact that this version of SQL Server works on everything from a notebook to a 16-processor symmetrical multiprocessor machine.

SQL Server 6.x database administrators will find a cornucopia of features aimed at easing their jobs. Many of the configuration tasks and settings that previously required manual adjustment are now more automated. For example, this version dynamically allocates and decreases memory resources on the fly. This is a big improvement over manual tweaking. Microsoft also has added utilities for improved backup and restore functionality.

I was most impressed with the improvements that were made to the replication capabilities in SQL Server 7.0. This version also supports update replication, which allows replicated data to be modified at multiple locations.

This feature will make SQL Server 7.0 particularly useful at distributed sites.

The tools included with this version were easy to use, and I especially liked the added support for Internet replication.

For accessing and manipulating various data types, SQL Server 7.0 includes support for OLE DB. In addition, access support has been updated via a new driver that is ODBC 3.0-compatible. (A 3.5 beta version of a ODBC driver manager is included.)

Developers will like the improved Transact-SQL enhancements, the monitoring and debugging capabilities, and the integration with other Microsoft development tools, such as Visual C++.

It is still early in the beta process, but if this release is an accurate indication - and I think it is SQL Server 7.0 is about to become a powerhouse in the enterprise database league. This featurepacked platform should appeal both to current customers and to those seeking an easy to use, easyto-implement database solution. SQL Server 7.0 looks like a team bound for the play-offs.

View Image - FORMERLY KNOWN AS ISQL/w, the SQL Server Query Analyzer includes a plan tab that shows detailed information on every stage of query execution, optimization, and index usage.

FORMERLY KNOWN AS ISQL/w, the SQL Server Query Analyzer includes a plan tab that shows detailed information on every stage of query execution, optimization, and index usage.

AuthorAffiliation

When not tackling databases, Maggie Biggs likes Wisconsin cheese, even though she didn't bet on the Packers. Send her your favorite fondue recipes to maggie_biggs@ infoworld.com.

Copyright InfoWorld Publications, Inc. Jan 26, 1998