Robert Crawford has 29 years of experience as a systems programmer. He currently is an operations architect responsible...
for setting the mainframe strategy and direction for a large insurance company in Texas. Here, he lays out his reasons for staying the course with mainframe computers, and tells how to go about a mainframe makeover as business needs change.
SearchCIO.com: With mainframe computers, the conversation is often about moving applications off of them, but are you saying enterprises are investing in new mainframes?
Crawford: We just recently signed a deal for new mainframes for the next five years. It's part of our strategic planning. I think a lot of big companies are continuing to invest in mainframes. They're not going away any time soon; and I don't think big companies that are heavily invested in them will be migrating away from them any time soon either.
Why did your company invest in new mainframes?
Our main moneymaking products are supported by mainframes. The applications for what we do -- our core businesses -- are running on the mainframes. I think that's true of a lot of other companies too. Some of it you could say is inertia: Yes, it would cost a lot to get off of them and it would be a huge effort. But when you look at economies of scale and of maintaining them, it's competitive with the distributed world.
Mainframes can run at pretty high utilization rates of 80% to 90% without breaking a sweat.
How is it possible to repurpose mainframes as your business needs change?
They're not necessarily being repurposed as much as [they're] growing into the new world. When I started in the '80s, most of the mainframe computing was green-screen, text-based applications.
As people have adopted the Internet for their customers and moved toward Web-based applications for employees, instead of being the presentation layer, [the mainframe has become] the back end that does all the heavy lifting: database operations and running through the business logic and doing the batch in the background. So, really, it's changing in that there are new interfaces being used to get to the data that's on it.
Business intelligence (BI) is one example of the need for data to sit on the mainframe. Are you building new interfaces for your mainframe so that more people across your organization can access the information on it?
Our transition is pretty much complete, going from text-based screens to Web-based app [interfaces]. Web-based apps are 95% of what the mainframe now supports. There has to be a distinction between operational systems and things like BI.
BI data used to be on the mainframe, but there are a lot of jobs and processes that extract that data from the mainframe and move it to other systems where the data analysis tools are friendlier for the people who have to do the analysis. The operational systems and the systems we present to customers are becoming Web-based. But there are other things that do get done on a distributed platform, like business intelligence and big data, because the distributed platforms are better at doing things like that.
What are the benefits of staying with a mainframe strategy?
More on boosting mainframe performance
Revitalize mainframe storage performance with VSAM
Using Agile development to boost mainframe performance
One is security. Mainframes are still a very secure environment. As far as we know, no one's written a virus that can get to the mainframe. Another one is economies of scale. In most shops, a small number of systems programmers are supporting a large amount of computing power. Just because of the monolithic structure of the mainframe, it makes it easier to maintain with a small staff. A lot of companies have [had] apps running on mainframes for the last 20 to 30 years, so they have quite a lot invested in it. A lot of competitive advantage is roped into the code that's in the mainframe, so they'd like to maintain it.
And really, it's a high-performing environment as far as database and transaction systems go. It can do an awful lot of work for an awful lot of people, where you have thousands of users going to one, concentrated space to do all this processing. That's also where the economies of scale come in. Then there's the fact that mainframes can run at pretty high utilization rates of 80% to 90% without breaking a sweat. That makes more efficient use of the hardware. There's high process intensity in a little space, so it goes a little easier on the data center floor space.
Read part two of this interview, "Mainframe migration: Start with a clean slate."
Let us know what you think about the story; email Christina Torode, Executive Editor.