Microsoft users will be interested in many of the technology features in this release, particularly the security enhancements. However, Microsoft's multiple themes and positioning messages for Yukon have confused customers regarding the guiding strategy behind the software, and indicate that Microsoft isn't sure what it is working toward. Microsoft has failed to deliver any major or breakthrough architectural advances with Yukon. Yukon is a montage release, with a number or scalability, availability and security features -- but not enough to enable it to fully compete with IBM and Oracle at the very high end of the DBMS market (2,000 or more concurrent users and multi-terabyte DBMSs) -- and a number of "hooks" to support Microsoft's applications and Extensible Markup Language (XML), but not enough to change the role of SQL Server in the DBMS market.
What's in and What's out?
According to Microsoft, Yukon is planned to include:
- Tighter integration with the .NET Framework (for example, user-defined aggregates, data types and functions, and integration with the Common Language Runtime)
- Additional support for XML (for example, XML datatype, XQuery, XML, data manipulation language and XQuery Designer)
- Enhancements for availability (for example, snapshots and fast recovery)
- Scalability enhancements (for example, table range partitioning, support for nonuniform memory access and parallel database consistency checker)
- Performance tuning and maintenance enhancements (for example, database tuning advisor and version-control support)
- Security (for example, fine-grained administration rights, and separation of users and schema)
- Business intelligence enhancements -- including extraction, transformation and loading (Data Transformation Services), dimensional modelling and online analytical processing (Analysis Services), reporting (Reporting Services) -- and development/management tools to facilitate the use of these features
According to Microsoft, Yukon is not planned to include:
- Clustering within the DBMS (shared disk or shared nothing)
- Hash partitioning
- Yukon as the unifying storage engine within Microsoft applications
- File system support within the DBMS
Timing Is Everything
Microsoft is under pressure to deliver something to its customers, who have been paying for its Software Assurance (SA) maintenance program for the past four years. If Microsoft delivers Yukon before customers' SA agreements expire, the customers will have at least justified (or just broken even on) the investment in their SA programs. However, customers will need to renew their SA agreements to get future releases.
If Yukon is released after customers' SA agreements expire, and they have not renewed their SA agreements, customers will have little value to show for their four-year investments. If they want Yukon, they will need to renew their SA agreements or purchase the product. The self-generated pressure is as much a factor driving Microsoft's delivery dates as completing the feature set. Furthermore, Microsoft is at serious risk of degrading SQL Server's market perception as a "total lower cost" alternative to Oracle and DB2.
We believe that many customers may consider not renewing their SA contracts because of a lack of perceived value. Enterprises should calculate the cost/benefit of an SQL Server SA agreement based on their business requirements. Enterprises acquiring or renewing SQL Server SA agreements should not build a return on investment case based on receiving a new version of SQL Server. Although there is no guarantee in an SA agreement that new versions of software will be delivered during the contract period, enterprises that find their SQL Server SA agreements have lapsed should try to negotiate with Microsoft for discounted renewal rates.
Key Issues
Which product and service vendors will become leaders in database management and administration?
How will database management technologies and standards evolve?
