Old and Getting Older

[Crossposted to The Mainframe Blog.] Part of my challenge in writing this blog is to provide enough meaningful information about why I'm in Japan without giving away sensitive IBM details. Please bear with me if I seem a little abstract in this update, but that's the reason.

Several days ago I mentioned that it's harmful to your business if you allow your IT infrastructure, including your mainframe, to deteriorate. Inertia can often be the most costly decision of all. One simple situation that everyone can understand is if a person is bleeding. Of course that person would seek a doctor immediately, understanding there will be costs. But the costs of inaction are far more dire: death is much more expensive.

The analogy is imperfect but valid. Let your mainframe age too much — get behind in your upgrades, hang onto old hardware too long, rely on unsupported middleware — and your business will suffer. You will pay more for software, maintenance, and support (if you can get it) every month. You will buy demonstrably more expensive computing solutions as your businesspeople desperately try to deliver new function, bypassing what you've neglected. You will bleed talented employees who will leave to work for a more progressive company. You will compete against businesses who have better technology and thus superior business functionality. You will actually experience more planned and unplanned outages with longer recovery times — newer hardware and software have substantial improvements in those areas. And you will risk compromising your customers' privacy because newer systems have better encryption facilities than older systems. Already one corporation which couldn't protect customers' privacy went out of business.

I guess it's no secret that the most technologically advanced industrialized country in the world, Japan, is struggling with this issue of aging IT infrastructure spectacularly and uniquely.

SOA (Service Oriented Architecture) is part of the answer. SOA fosters reuse of IT assets so that your technical infrastructure can respond much more quickly and easily to changing business demands. It's also one of the few ways businesses and governments will be able to reduce their IT expenses over the long term while actually improving their capabilities. SOAs will only be successful if the mainframe provides services (the S in SOA) and hosts at least the major part of the Enterprise Service Bus (ESB). (The ESB is the core foundation of any moderately sized SOA.) Quite simply, if your mainframe is not the key ingredient in your SOA then you really don't have an SOA strategy.

Mainframes are as youthful as you want them to be, and they deliver the highest qualities of service for the lowest total costs per transaction. SOA means you can exploit your mainframe's abilities for maximum business advantage. You can bet at least one of your competitors has already figured out these facts.