A few months ago, I wrote about the possible outcomes of Sun's product suite. Since that time, the merger has been delayed and Oracle has made public announcements that it intends to keep Sun's hardware business to compete with the likes of IBM.
Oracle also announced Exadata v2, a new high-performance database appliance based on Sun hardware. This has generated additional confusion regarding the future of Sun-Oracle. What do these changes mean? What can Sun customers read into these rumblings?
Oracle's server hardware contradictions
On September 10, Oracle posted an second version of the Oracle Exadata high-performance appliance. While Exadata is built on Sun hardware (as opposed to Exadata v1, which was built on HP hardware), it does not incorporate any of the promises outlined in the September 10 advertisement to Sun customers. That's right: Exadata v2 is built on Sun's Intel Xeon 5500 series-based servers running Oracle Enterprise Linux. No SPARC or Solaris in this appliance.
While this may appear as a cause for confusion, if we think back in time, we realize that the Exadata v2 appliance must have been well in flight at the time of the Oracle-Sun merger announcement earlier this year. Integrating software and hardware to create a product like this for enterprise deployment and support requires a long period of development, testing and support preparation. Making course changes to integrate with Solaris and SPARC would have resulted in significant product delays.
Therefore, the Exadata v2 announcement cannot be reliably used to predict what the merged Oracle-Sun company will do with its hardware.
SPARC and Solaris customers speak out
I've had the opportunity to talk with a number of customers since my June article on the subject of SPARC and Solaris deployments, and I've learned some things that I'd like to share.
- Entry-level and midrange SPARC servers cannot compete with commodity Intel and AMD equivalents on a price/performance basis.
- High-range, scale-up SPARC servers do not (yet) have a compelling commodity Intel or AMD replacement that can scale. I'm talking about 64-core and greater machines used for a very limited class of scale-up applications (mostly database).
- Nearly all entry-level and midrange SPARC customers have migrated or are in some stage of migration from SPARC to Intel- or AMD-based commodity servers.
Given this information, we can see more clearly the best roadmap for Sun SPARC hardware. It simply boils down to what the application's needs are.
IF the application's scalability requirements can be met with Intel- or AMD-based commodity servers,
THEN migrate from entry-level or midrange SPARC servers to commodity AMD or Intel servers.
ELSE, maintain or upgrade high-end SPARC server to support the application's scalability requirements until a future time when commodity servers can meet the application's scale-up requirements (two to five years, depending on needs).
Only a handful of enterprise applications will fall into this category, but nearly all of them will be dependent upon Oracle's cornerstone product: the Oracle 11g database.
Oracle's probable Sun hardware plans
Based on this information plus the need for Oracle to protect its high-end database revenue, it does make sense for Oracle to invest in the high-end SPARC- and Solaris-based servers for those application/database configurations that require scale-up hardware environments.
While these cases are few in comparison with other workloads, they tend to be mission critical for the customer, and high-margin sales for Oracle and Sun. Additionally, it is with these high-end workloads that Oracle will need to compete with IBM in order to ward off the threat of IBM's DB2 database product.
Expect the entry-level and mid-range SPARC hardware offerings to be discontinued or pushed to a partner such as Fujitsu.
The EU Factor
The European Union's probe into the database market (given that Oracle will also own the MySQL database products after the merger) has placed additional pressure on Sun and Oracle to do something. This probe was announced September 3 and is undoubtedly the cause of the Oracle open advertisement to Sun customers about its SPARC and Solaris plans. Further delays on the merger only grant Sun's hardware competitors more time to harvest fearful and weary Sun customers. Oracle's biggest threat is IBM -- the competitor specifically identified in the Oracle advertisement.
Unfortunately, we have longer to wait until we hear the definitive plans for Sun's assets from Oracle.
ABOUT THE AUTHOR: Richard Jones is vice president and service director for Data Center Strategies at Midvale, Utah-based Burton Group. He can be reached at email@example.com.
What did you think of this feature? Write to SearchDataCenter.com's Matt Stansberry about your data center concerns at firstname.lastname@example.org.
This was first published in September 2009