Infrastructure From Application Vendors: What CIOs Want and Need

Enterprise application vendors like SAP, Oracle, PeopleSoft, and Siebel, are on a collision course with infrastructure vendors like IBM, Microsoft, Oracle, and BEA. Application vendors are offering infrastructure software that will eventually compete directly with software traditionally purchased from infrastructure providers, namely portal, integration, Business Process Management (BPM), and in some cases, application server technology.

The Bottom Line: CIOs are likely to resist application vendors’ attempts to sell infrastructure technologies. The implementation is not quite what CIOs are looking for, despite looking sound in theory. Still, the continuing clash between infrastructure and application vendors will have a momentous effect on the future of the enterprise technology market.

What It Means: While the infrastructure moves by application vendors promise to exploit existing investments in application software, CIOs are skeptical of the vendors’ ability to deliver truly open and interoperable infrastructure software.

According to the many CIOs, application vendors are claiming that their newest infrastructure technology can help them in several ways by:

• Leveraging their existing application investment
• Interoperating with other applications and infrastructure
• Reducing cost of ownership and acting as logical, evolutionary, and safe bets

However, CIOs interpret the story differently, sensing ulterior motives:

• Selling them a fix for the application vendor’s integration problem
• Locking them in as clients and eliminating future choices
• Getting greedy, seeking new revenue sources, with infrastructure software becoming the field on which to fight for a larger slice of the enterprise pie

At our annual Strategy 21 conference this week, we compared and contrasted two enterprise application vendors’--SAP’s and Siebel’s--approaches to the infrastructure market.

SAP’s NetWeaver

SAP is moving in the right direction by offering functionality requested from CIOs, as documented by AMR Research in 2001. The new technology includes frameworks for portals, analytics, process and data management, integration, and composite applications.

Still, it’s not exactly what CIOs want, for a couple of reasons:

• It appears to be an SAP-only stack, rather than an open architecture that can be used outside of the SAP application ecosystem.
• It seems doubtful that any other major vendor will ever port its application to run on SAP NetWeaver for fear of being locked into SAP-only accounts or being subsumed into the SAP application.

Siebel Universal Application Network (UAN)

In contrast, the Siebel Universal Application Network (UAN) tackles the problem uniquely, offering Business Integration Applications (BIAs) that are built to deploy across multiple integration servers. What’s most appealing is that many application vendors could share the approach, each producing its own BIAs and publishing them to work universally with one another. A network of integration servers from vendors like webMethods, TIBCO, IBM, SeeBeyond, BEA, Microsoft, and Vitria could then coordinate the BIAs.

However, the Siebel approach also has its weaknesses, mostly because it relies on an as-yet unachieved adoption of standards:

• A method for authoring processes
• A standard repository for storing process models
• Shared standards for integration servers

The Takeaway

Overcoming these problems will require a consortium of vendors, both application and infrastructure, to cooperate. In the best case, this could be a win-win-win scenario for all parties: application vendors, infrastructure vendors, and CIOs. In the meantime, there are steps companies can take now as the hard sell comes their way.

Recommendations:

• CIOs should resist purchasing infrastructure technology from application vendors until the vendor can prove its architecture can be deployed across a variety of enterprise applications.
• Since it’s not a problem that can be isolated to infrastructure, don’t accept Web services as the simple answer. The solution will require cooperation from infrastructure and application vendors.
• Best-of-breed vendors will not likely cooperate with the large suite vendors’ proprietary approaches to infrastructure. They will have to band together and find a better way to interoperate. Siebel’s UAN and BIA approach is among the possibilities.

Conclusion: This impending collision between enterprise infrastructure and applications represents a multibillion-dollar inflection point, with the future growth of all software providers dependent on their chosen strategy. A successful Oracle bid to purchase PeopleSoft would weigh heavily on what happens in the next three years, dividing the large and the small and mandating a radically different approach to interoperability.

AMR Research originally published this article on 11 February 2004.