You know what? Oracle doesn’t need an OS. When Larry Ellison mused about coming up with his own Linux distribution recently, he was dreaming about his company’s stack. It’s an ego-stroking fantasy not unlike those triple threats in the entertainment industry.
These include writer/director/actors, producer/singer/songwriters or even singer/actor/authors. The difference is that in show business, taking on additional duties is a way for successful artists to spread their wings, even if they face some difficulty being taken seriously in their new pursuits. In the IT business, vendors are giving themselves extra credits because if they don’t, they worry no one will take them seriously at all.
The notion of “core competency” used to be a big deal in IT circles, but today you only hear the term used in reference to customers, who are urged to hand over more of their functions to outside providers. A business applications vendor, on the other hand, can’t respond to client inquiries about management tools by saying that it falls outside its core competency. That would mean they are literally incompetent to handle that client’s evolving IT requirements. After years of high-tech firms ignoring, selling off or quietly killing the less valuable portions of their software product portfolio, the stack is back, and in a big way.
Although challenges around integration are driving vendors to build the complete stack, it’s not the only issue. It’s a reaction against the “best of breed” theory, in which IT managers could pick and choose the pieces that run their infrastructure from among the highest-quality options. No one would dare say customers don’t deserve choice, but the best-of-breed model simply doesn’t provide long-term growth prospects. The best that can happen is you excel in one area, such as CRM, until you’ve tapped out your potential prospects. That’s about the time when people like Tom Siebel start taking Larry Ellison’s phone calls again.
But there’s more to it than that. The truth is, “best of breed” never existed. If it did, creating service oriented architectures wouldn’t be such hard work, and the open standards everyone supposedly builds upon would be enough for CIOs to swap products in and out of their network with ease. That isn’t what’s happening in the enterprise today. Certainly there is “better breed.” Open source startups such as MySQL wouldn’t be making an impact if Oracle and IBM were completely satisfying the database market’s needs.
The tough industry economics that have led to massive consolidation in software means some vendors are assisted in their stack-building efforts. But at this point, would it really make sense for HP to launch something to compete with DB2 or SQL Server? Aren’t companies such as Red Hat, Novell and BEA better off partnering than coming up with their own middleware to fend off Fusion or NetWeaver? If Oracle Linux were available today, what kind of investments would Ellison have to make to successfully supplant Microsoft Windows’ position in Oracle’s installed base?
When customers talk about wanting “one throat to choke,” it doesn’t mean they mind seeing other brand names in their data centre. It means they want leaders in the software industry that will broker meaningful relationships among their peers so that problems can be identified, diagnosed and solved by approaching any vendor within their IT supply chain. There is no point in trying to replace best of breed with a so-so stack.
Shane Schick is the editor of ITBusiness.ca.