A year ago, I met with the Chief Strategy Officer (CSO) from one of the largest open system hardware companies. Surprisingly, the CSO's interests turned our discussion onto the future of the mainframe. For those that may not be aware, I had formerly led CA Technologies' mainframe business unit for 3 years and led the launch of our Mainframe 2.0 strategy. I assume given my background, the CSO had sought me out to pick a fight and convince me on the looming death of the mainframe.
He initially conceded that until very recently, open system hardware had certain advantages over mainframes (price performance being the most often stated), but there were always show stopper technical deficiencies that prevented any large scale market defection from the mainframe platform. Today however, he argued, the latest advances in open systems had surpassed the mainframe on every technical measure by a highly compelling margin and that the resulting economic value would finally bring old "big iron" down.
I'm not the one to judge the validity of the CSO's "speeds and feeds" argument, but he was missing the most critical point. The mainframe market collectively has invested 100's of billions in proprietary application development on the mainframe platform. Moreover, that huge investment still runs at a level of reliability, availability, and serviceability that remains unique to the platform and which companies have built their business models to depend on. It doesn't just work...it works amazingly well. With all the critical challenges that large enterprise CIO's face at the moment, a mainframe conversion is a high risk/low reward proposition that isn't even worth the time to consider.
I then turned the tables on the CSO and made the point that I believe the most threatened platform in enterprise data centers is not the mainframe, but rather open systems that will succumb to the inevitable gravitational forces from the cloud (gravity from a cloud would require an awfully big cloud ). I politely cautioned him that he should focus himself less on the mainframe market and be most concerned about his own companies future.
My reason for bringing this up was based on a article that I read today form Forrester (Forrester: Cloud is not the future of IT. Long live mainframes? See Article Here). The article implies that enterprise data center's last remaining platform will be the mainframe interconnected to an array of cloud services. Can you imagine that?
An enterprise IT model of "big iron" and "big clouds" means "big problems" for those vendors depending on open system sales to enterprise accounts. Long live the mainframe? You betcha!