October 6, 2011
OK, out of respect for your time, now that I’ve caught you with a title that promises some drama I’ll cut to the chase and tell you that I definitely lean toward the former. Having spent a couple of days here at Oracle Open World poking around the various flavors of Engineered Systems, including the established Exadata and Exalogic along with the new SPARC Super Cluster (all of a week old) and the newly announced Exalytic system for big data analytics, I am pretty convinced that they represent an intelligent and modular set of optimized platforms for specific workloads. In addition to being modular, they give me the strong impression of a “composable” architecture – the various elements of processing nodes, Oracle storage nodes, ZFS file nodes and other components can clearly be recombined over time as customer requirements dictate, either as standard products or as custom configurations.
The notion of workload-specific integration of systems is not unique. As my colleague Frank Gillett has clearly outlined in “The Age of Computing Diversity” and I have discussed in several blog posts on converged infrastructure offerings and even entire data centers designed around specific workloads, there is value in tailoring systems to workloads if the economics are clearly understood. The thing about Oracle’s systems is not so much the technology, although it is elegant. IBM’s Netezza BI appliances offer similar architectural blending of compute nodes, storage processors and even very specialized engines to do low-level database related functions. The real aspect of these systems that should send frissons of fear through the corporate nervous systems of competitors is the workload – Oracle applications, especially its eponymous database system, represent the “killer app” for UNIX, and its enterprise business applications are also strong contenders in aggregate. Oracle’s database may, depending on which self-serving estimate you believe, represent between 25% and 50% of the workload (defined by me as the primary workload and the reason the system was purchased) for the still approximately $20 billion RISC/UNIX systems market. With Oracle having put a severe crimp in HP’s UNIX business and with an installed base of tens of thousands of legacy SPARC systems, Oracle is in a position to move these things like hotcakes into an installed base that is very hungry for capacity upgrades and consolidation, with the only credible competitor for RISC/UNIX Oracle database workloads now being IBM. Additionally, Oracle will get a share of other adjacent workloads such as ERP, BI and Java-based private clouds. All in all, not a bad position to be in only a couple of years after competitors had counted you down and out as a serious systems competitor.
Competitors HP and IBM are fighting back with an impressive rollout of their own application-specific bundles, but with the exception of SAP, which will likely remain a defensible stronghold for both HP and IBM (although Oracle will get some share of this), it appears that Oracle has a very strong position with its Engineered Systems for the RISC portion of the core Oracle workloads as long as they can deliver on their price-performance promises. I&O executives are in the end pragmatists, and despite their generally negative perceptions of Oracle’s business practices (especially current HP-UX customers), they will continue to do business with Oracle, and will purchase Oracle licenses. Many will be on x86 Linux, of which HP will get a disproportionate share, and the remainder will be essentially split between Oracle and IBM RISC systems. HP will strengthen an already solid relationship with SAP, fueled in part by their mutual loathing of Oracle, but even there will face competition from Oracle and IBM.