By James Kobielus
Last fall, Oracle CEO Larry Ellison announced that his company was getting into the hardware business, but I think he misspoke. At that time, he was referring to the new HP Oracle Database Machine with Exadata Storage, a high-end data warehousing (DW) appliance that incorporated hardware from his partner, as well as intelligent storage software technology from that partner--and even had the partner’s name first in the product name. If that was the criterion for “getting into the hardware business”--i.e., running on someone else’s hardware--then every software vendor on earth is in the hardware business, by my reckoning.
But today’s Oracle announcement is the real deal. Oracle is acquiring longtime partner Sun Microsystems, putting the software powerhouse fully into the hardware business--and hitting the DW industry like an earthquake. I’ll let my Forrester colleagues blog on the other implications of this deal--for the open source, Java, middleware, SOA, and other markets that Sun is in--and give you a few quick thoughts on the deal’s implications for the DW market.
For starters, this deal will give Oracle the ability to engineer a completely integrated DW appliance composed of all Oracle components, including hardware and software. Now Oracle will be able to take on Teradata and IBM--both of which have long offered their own integrated solutions--more aggressively with high-performance DW offerings. Just as important, Oracle will be able to leverage Sun’s manufacturing scale economies to bring its all-Oracle DW appliances below the $25K-per-terabyte threshold needed for penetration into the midmarket.
Also, Oracle will now have another widely adopted transactional database, the open-source MySQL, that it can--and should--consider tweaking and packaging on an DW appliance. To the extent that Oracle gives customers a choice of DBMSs on a DW appliance platform, it can gain a differentiator that Teradata, IBM, Microsoft, Sybase, and Netezza lack (you have to go to a startup such as Dataupia for multi-DBMS choice on an appliance). Many information managers prefer to stick with their existing DBMSs when building a DW, and prefer to implement that DW on an appliance to take advantage of its out-of-box balanced configuration of CPU, memory, storage, and I/O.
Furthermore, Oracle is acquiring a hardware and operating system vendor that has long been one of the primary platforms on which its own DW/DBMSs, middleware, and tools have been deployed. This acquisition can only be welcome news for joint Oracle-Sun DW customers who have worried about Sun’s solvency for some time now and began to sweat serious bullets when IBM failed to emerge as a white knight. For many Sun customers, an Oracle-powered DW platform will now look like a safer bet than ever.
Of course, there are clear risks in this pending acquisition.
First, a combined Oracle/Sun sows uncertainty among the DW appliance vendors--such as Greenplum and ParAccel--who have partnered with Sun and now find themselves in earnest “co-opetition” with full-competitor (and then some) Oracle.
Second, Oracle’s other DW appliance hardware partners--including HP, IBM, and EMC/Dell--must be concerned that Oracle will now shift focus away from their respective appliance products in favor of those it builds with its own Sun hardware group.
And finally, Oracle’s acquisition of Sun--and possible future development of a MySQL DW appliance--may discourage customers from considering third-party DW appliances, such as from Kickfire--that build on MySQL. If that happens, and a market for non-Oracle-branded MySQL DW appliances never takes root, Oracle will be denying its MySQL customers the choice that Oracle Database customers already enjoy. Currently, Oracle Optimzed Warehouse customers can deploy that enterprise DBMS as a DW on their choice of Sun, HP, IBM, and EMC/Dell platforms.
Let’s hope that Oracle makes the most of its pending Sun acquisition. Ellison either misspoke last fall, or was speaking prophecy. Like most DW vendors, Oracle’s destiny is to grow ever more hardware-dependent for its long-term scalability, performance, and optimization story.
For more information on Oracle's Aquistion of Sun click here.
Agree with your analysis that this deal will push Oracle towards data warehousing appliances. With the HP partnership, they had already recognized that appliances were the way to go to get data warehouse performance. Now, as you say, the have the opportunity to deliver something additional, both on their high-end database, Oracle, and their low-end database, MySQL.
At Kickfire (www.kickfire.com), we see this as all good news. Kickfire's chip makes MySQL run 10-100X faster for data warehousing. Our patented chip enables query parallelization on MySQL, which as you know, is critical for data warehousing performance. We are the only vendor that can do this and therefore deliver this kind of performance on MySQL.
The chip comes packaged as an appliance today. Inside the appliance, the SQL chip is plugged into a Linux base server from Intel through a standard PCIe connection, just as you plug in an NVIDIA card. What if you plugged the SQL chip into say a Sun Fire server? You now have a server that can run MySQL 10-100X faster than the competition. I think some people will find that interesting.