Next G Business Applications: game on IBM and Oracle, and so long HP

When Larry Ellison throws down the gauntlet of transaction processing benchmarks and offers to pay "any company $10 million if [their] Oracle database application doesn't run at least twice as fast on SUN hardware as on IBM's fastest computer", there is no doubt that a rivalry exists between the Red Army of Redwood City and the Blue Suits of Armonk.

What began in the 1980's as a contest at the database layer has progressed over the last quarter of a century ever upward through an increasingly complex software stack. Regardless of how different these two giants of the ICT industry may seem on the surface, the truth is that the real battle has always been about who runs the world's enterprise level business applications. Next year is shaping up as the inflection point from which true supremacy in this arena will be decided.



At first the suggestion that IBM and Oracle will compete head-to-head on "packaged" or "commercial off-the-shelf" business or enterprise applications may appear unlikely; after all the two companies appear to have extremely different strategic directions. On the one hand IBM has acquired and expanded its services business while evolving its software infrastructure capability. It has continued to divest itself of almost all its hardware businesses and has previously stated it was not in the business of building packaged applications. In contrast, Oracle has acquired some of the largest packaged application vendor brands on the market, from JD Edwards to Siebel, encompassing everything from traditional ERP to CRM. They have continued to support these seemingly overlapping solutions, avoided any move into the high-end or large-scale services area, and more recently began the process of acquiring a hardware vendor.

However, it is within these two very different approaches that Longhaus perceives the same endgame; the creation of widely applicable, highly configurable, service-oriented business application assets that can be deployed to an open and standards based middleware stack. It is in pursuing this end-game that IBM is relying on the increasing production of business level software products based on client implementations using the process and information models within the Industry Frameworks that have made GBS such a successful consulting and services business. Oracle is relying on the production of best-of-breed components based on the process and information model patterns found within its extensive portfolio of near identical packaged applications. In both cases the secret is in harvesting the world's next generation of enterprise level business applications from existing pools of intellectual property.

When placed side-by-side (see Figure 1) it is fundamentally the source of this intellectual property that differs. For IBM and Oracle it is the very different paths travelled from 1980's database vendors to 2000's mega-vendors that has led them to those sources. Even the acquisition of business intelligence tools, including Cognos by IBM and Hyperion by Oracle, to provide a more information driven user experience re-enforces the similarity in the final product offerings that will emerge as the second decade of the new century progresses.

So are these next generation business applications like the packaged or commercial-off-the-shelf applications of the past? Yes and no. They are certainly commercial offerings with associated licensing and maintenance fees for in-house or on-premise deployments. Or they are subscription fees when made available as-a-service. But they are certainly not stand-alone like the packaged applications of old requiring instead that a foundation of service-oriented middleware be in-place before they can be deployed or executed. One way to conceptualise these emerging offerings that IBM Software Group call "Industry Solutions" and that Oracle Products Group call "Fusion Applications" is the clothing, or business suit analogy. At one end of the spectrum organisations can purchase their applications fully finished; the off-the rack or ready-to-wear solution in set sizes.

At the other extreme is the in-house or custom developed solution; the equivalent of a bespoke tailor made solution that will fit the organisation like a glove once finished, but cost a whole lot more than off-the-rack. What IBM and Oracle's strategies for business applications represent is something more akin to "made to measure". In this model the core solution is pre-made to reduce "fitting" (think implementation) time, but the scope for adjustment is much wider than an off-the-rack solution. This scope for adjustment allows IBM and Oracle to address industry nuances and other subtle organisational needs to be included.

While the business application level strategies of IBM and Oracle are now starting to emerge with the obvious conclusion that they are on a collision course, there remains one problem. Neither IBM nor Oracle is as yet able to fulfil the vision they have outlined. Sure, they both have highly mature, well integrated, open- and standards-based middleware, but middleware won't sell itself. It is the business application components that require these middleware platforms that offer the true business value.

In the case of IBM's Industry Software Solutions, while mature in some areas such as software for Telecommunications Service Providers, they are only just emerging in other areas such as Network Centric Options solutions for the Military and Defence. IBM's industry approach is also heavily focused on the middle office or line-of-business applications, leaving them wanting and reliant on SAP (or ironically even Oracle) when it comes to traditional ERP functionality.

As for Oracle, new versions of the current stable including Oracle eBusiness Suite, Siebel and PeopleSoft can deliver an approximation of next generation business applications, but the back-end architectures of these solutions still carry much of their original legacy designs. Oracle therefore is left in a position of having only demonstrated the Fusion Applications. They are a much needed offering but won't officially be available until next year.

The fact that neither IBM nor Oracle have fully realised a complete set of next generation business applications is by no means lost on either party. Larry Ellison's recent public remarks and quiet confidence are testament that the race is on and the announcements from both firms over next 12 months will be significant; each customer win will represent an important milestone. What remains to be seen is whether HP will leverage its new services arm's tight links with SAP to craft its own business application strategy. Unless they do, the enterprise level business application market of the next decade will remain a two horse race.