Attended what was billed as a smackdown between SAP and Salesforce at the Churchill Club's "Great Debate: The Future of Enterprise Software." The two belligerents were Hasso Plattner, co-founder of SAP, and Mark Benioff, found of Salesforce.com.

Overall, the discussion was supposed to focus on the potential of SaaS vs. on-premises; however, there was a lot of jostling between the two, with Benioff particularly pitching the virtues of Salesforce.

Nonetheless, the session was enlightening, and held some interesting lessons for you, the CIOs of the world.

Benioff pushed the notion of multi-tenancy hosted apps, and how that get critical updates immediately applied to every user organization. According to him, the Salesforce.com offering, which has been extended to be a general purpose platform, should be the basis for everyones' apps going forward -- including, he maintain, SAP. This is because Salesforce has such experience and insight about building highly reliable, scalable infrastructure.

Plattner, on the other hand, noted that SAP is hard at work creating an on-demand offering for those customers that choose to use SAP in that fashion. SAP has 2500 developers building the offering, and is planning to offer 2100 APIs to enable companies to customize the product; necessary because their customers insist that an on-demand product be capable of offering the same level of functionality as the on-premise version.

I have to say that I felt that Benioff's statements seemed to have a lot of fluff associated with them. While he implied that they are building out a full set of enterprise functionality, I don't understand how an end user can customize the product for their own purposes and still realize the benefits of multi-tenancy, including transparent upgrades. And whenever I hear someone maintain that I should build my infrastructure on their proprietary platform, I hear the phrase "lock-in" somewhere in the back of my mind. Just because I don't have to install the software on my own servers doesn't mean that lock-in doesn't limit my freedom all the same.

Regarding Plattner's on-demand discussion, it's hard to not see Clayton Chrisensen's Innovator's Dilemma infusing their efforts. SAP has built an incredible business offering the largest companies in the world the ability to automate their operations with individually tweaked installations. As it creates its on-demand offering, the expectations of those customers molds the offering, forcing it to enable the same level of complexity and customization, no matter how poorly suited for an on-demand architecture. I mean -- 2500 developers? 2100 interfaces?

Overall, however, I was struck by the underlying assumption that enterprise software, qua software, will continue as it has always been: expensive, complex, and sold by high-margin companies to locked-in lower margin companies. I did a study for a client once on Salesforce TCO and concluded that, while the payments are spread over a longer period of time, its total cost is equivalent to the packaged alternatives.