Oracle is the most interesting of the ‘whale’ companies that I follow. The firm to me is really an M&A company with an enormous software franchise and now a hardware portfolio that is very credible. Ellison and company have an unbelievable track record of making good calls and even the vocal Microsoft bashing of the mid-90’s and the prediction of a thin client Tsunami, while somewhat off-base is coming to fruition more than a decade later.
The most interesing thing to me is Oracle’s move to invest in IP – or what we refer to as the enterprise stack. Oracle’s stack is the most robust in the business. The only piece it lacks is networking and it really doesn’t need networking. It’s services business could be deeper to compete with IBM and HP and I could see Oracle making a move there but the company is incredibly well-positioned to continue to suck the air out of the enterprise market space. The center of Oracle’s universe is the database but increasingly the company is recognizing that it needs to have offerings that don’t customers to lock into its database. Oracle has an aggressive middleware strategy that it can use to pivot customers into its larger stack offerings, including applications, database and hardware.
From a hardware perspective it seems to me that Oracle is happy to commoditize that layer of the stack and continue to drive more value into the application layer. Oracle Automatic Storage Management (ASM) is a good example. Customers I talked to love the feature and it’s a no-brainer for them to adopt. Increasingly, Oracle and other ISVs will place application function into their stacks and go after the domain of infrastructure players. Because Oracle’s margins are so much more obscene on software, the firm is happy shifting function up the stack. Application recovery is another area where Wikibon members are emphasizing placing function into the software stack. Many members have explicitly indicated they’re moving away from hardware-based data replication (e.g. SRDF) and more toward an integrated software model. The advantages are it: 1) simplifies the management of infrastructure and 2) provides cleaner, faster and often more reliable recovery.
The challenge users have with this approach is that the service (in this case remote replication) is very application specific and can’t be shared broadly across the application portfolio. Hardware assets are more fungible and can be shared widely across applications and as such, at scale, a shared hardware approach will generally be more cost effective. The same phenomenon is true for integrated data warehouse engines such as Exadata. Customers love the integration approach and Ellison has often alluded to the iPhone as a good reference model for the integrated data center. However this approach will be generally more expensive (albeit often it will drive more business value for users). Nonetheless, it appears that the vast majority of the marketplace is moving toward database appliances that are integrated and Oracle has likely made another good call with Exadata. Indications are that Exadata and products like it are exploding with many customers buying multiple systems for deployment.
The bottom line for me is Oracle is not so quietly carving out a place as a dominant player in the data center and in my opinion is one of the early favorites in the race to the next generation data center. It’s virtualization strategy (OVM) is behind VMware’s however as I’ve previously written, in my view, all roads to virtualization will lead through Oracle eventually because of its application and database dominance, I can’t see massive shifts to virtualizing mission critical apps without the blessings of application heads and those professionals are very tight with Oracle.
Oracle OpenWorld is coming up next month in SFO and I think along with VMware it’s become one of the most important events in the business. I look forward to following the developments leading up to and at the show.