Last month, I had the pleasure of speaking at a meeting of the Cleveland, OH-based IT user group, E-STORM. E-STORM is a pure grassroots effort that brings together IT professionals from every level of Ohio companies—from managers to storage administrators, and from mainframers to Web server jocks.
The amazing thing to see was how well everyone got along. Advocates of x86 virtualization strategies were comparing notes with z/OS-philes who talked about PRISM, LPARs, and Sysplex. Mainframe backup mavens were sharing war stories with distributed system data protectors and explaining how to create backups as batch processes that would complete within the operational windows. Dogs and cats living together . . . .
After my talk, a young guy came up to thank me for including mainframes in my discussion as a relevant technology. He said that he was tiring (as am I) of both the mainframe is dead dogma, and of the mainframe is not dead stories he occasionally reads in the trades. He focused on mainframes in his university education and cultivated a set of skills in mainframe systems programming that he took for granted as enabling him to put food on his family’s table for the rest of his career. A mainframe, after all, is just another platform, we agreed, well-suited to performing certain tasks from a price and performance perspective, just as x86 is suited to other jobs.
I wish the marketing folks for system vendors, both big and little iron, could have been at this meeting. They might have learned a few things, both about real-world technology and about the people who use the tools. The facts seem to be getting lost in their product messages.
Mainframe folk aren’t “cumudgeony grandpappies” prone to lengthy ruminations about the way they did things back then, at the dawn of the computer revolution. Sure, their jokes about punch cards, tape reels, and disk packs may sound a tad dated, but that’s because, like most IT folks, they get fed through a slot in the bottom of the door at work. Ask them about world politics, the Democratic primaries, or global warming, and they will give you some rather current and probative opinions.
Nor are the distributed system folk condescending, little upstarts who diss their elders by wearing iPod headsets when you try to talk to them. They might find certain buzz technologies, such as XML, SOAP, and Web services, kind of cool. But, once you get past the marketecture, you might just find you’re dealing with some rather impressive depth of knowledge on x86 processor extensions, LUN mapping techniques, and the vulnerability of RAID schemes as drive size capacities exponentially increase.
It isn’t us vs. them, and it probably never has been. The entire big iron vs. distributed systems war was manufactured by the vendors. In many ways, the current history of computing is about how two architectural models can be combined to get the best each world has to offer. I have yet to meet a distributed computing advocate who doesn’t wish he had the management toolsets that are built into mainframe architectures. By the same token, I’ve talked to very few mainframers who aren’t interested in improving their ability to rapidly adapt to changing business needs or play with cool widgets and gadgets that keep popping up in distributed systems.
The one thing all computing professionals seem to have in common is that they’re all geeks at heart. Witness all the IBM executives who were carted out in front of analysts and journalists in April to discuss “mashups” as a userempowering application development methodology. The idea of a mashup is to use discrete code building blocks to create new application software. Making the process userfriendly supposedly democratizes the application development process.
From where I’m sitting, mashups can be either a road to Hades that’s paved with good intentions (is giving application development to users really such a good idea?) or a cool remedy for the rift that has descended between the front and back offices in many companies. The latter will be the case only if we’ve learned the hard lessons gleaned from the move into a distributed computing world, a move that threw out the baby with the bathwater in terms of effective systems management. I’m sure hoping we have and the ESTORM experience suggested it is so.
If you’re still reading, I offer only this as a conclusion to this month’s column. Whenever you sense a vendor marketing brochure or advertisement is setting up either mainframe or x86 computing as a whipping boy, clip the ad and mail it back to the vendor in question. Attach a note that says simply and clearly, “Them is us.” Z