This latest BriefingsDirect podcast discussion tackles the high -- and often under-appreciated -- cost for many enterprises of doing nothing about aging, monolithic applications. Not making a choice about legacy mainframe and poorly utilized applications is, in effect, making a choice not to transform and modernize the applications and their supporting systems.
Not doing anything about aging IT essentially embraces an ongoing cost structure that helps prevent new spending for efficiency-gaining IT innovations. It’s a choice to suspend applications on ossified platforms and to make their reuse and integration difficult, complex, and costly.
Here to help us better understand the perils of continuing to do nothing about aging legacy and mainframe applications, we’re joined by four IT transformation experts from Hewlett-Packard (HP): Brad Hipps, product marketer for Application Lifecycle Management (ALM) and Applications Portfolio Software at HP; John Pickett from Enterprise Storage and Server marketing at HP; Paul Evans, worldwide marketing lead on Applications Transformation at HP, and Steve Woods, application transformation analyst and distinguished software engineer at HP Enterprise Services. The discussion is moderated by me, Dana Gardner, principal analyst at Interarbor Solutions.
Here are some excerpts:
Evans: What we’re seeing is that the cost of legacy systems and the cost of supporting the mainframe hasn’t changed in 12 months. What has changed is the available cash that companies have to spend on IT, as, over time, that cash amount may have either been frozen or is being reduced. That puts even more pressure on the IT department and the CIO in how to spend that money, where to spend that money, and how to ensure alignment between what the business wants to do and where the technology needs to go.
Our concern is that there is a cost of doing nothing. People eventually end up spending their whole IT budgets on maintenance and upgrades and virtually nothing on innovation.
At a time when competitiveness is needed more than it was a year ago, there has to be a shift in the way we spend our IT dollars and where we spend our IT dollars. That means looking at the legacy software environments and the underpinning infrastructure. It’s absolutely a necessity.
Woods: For years, the biggest hurdle was that most customers would say they didn’t really have to make a decision, because the [replacement] performance wasn’t there. The performance-reliability wasn't there. That is there now. There is really no excuse not to move because of performance-reliability issues.
What's changing today is the ability to look at a legacy source code. We have the tools now to look at the code and visualize it in ways that are very compelling.
What has also changed is the growth of architectural components, such as extract transform and load (ETL) tools, data integration tools, and reporting tools. When we look at a large body of, say, 10 million lines of COBOL and we find that three million lines of that code is doing reporting, or maybe two million is doing ETL work, we typically suggest they move that asymmetrically to a new platform that does not use handwritten code.
That’s really risk aversion -- doing it very incrementally with low intrusion, and that’s also where the best return on investment (ROI) is. ... These tools have matured so that we have the performance and we also have the tools to help them understand their legacy systems today.
Pickett: Typically, when we take a look at the high-end of applications that are going to be moving over and sitting on a legacy system, many times they’re sitting on a mainframe platform. With that, one of the things that have changed over the last several years is the functionality gap between what exists in the past 5 or 10 years ago in the mainframe. That gap has not only been closed, but, in some cases, open systems exceed what’s available on the mainframe.
It’s not only a matter of cost, but it’s also factoring in the power and cooling as well. Certainly, what we’ve seen is that the cost savings that can be applied on the infrastructure side are then applied back into modernizing the application.
Hipps: This term "agility" gets used so often that people tend to forget what it means. The reality of today’s modern organization -- and this is contrasted even from 5, certainly 10 years ago -- is that when we look at applications, they are everywhere. There has been an application explosion.
When we start talking about application transformation and we assign that trend to agility, what we’re acknowledging is that for the business to make any change today in the way it does business -- in any new market initiative, in any competitive threat it wants to respond to, there is going to be an application -- very likely "applications" plural.
The decisions that you're going to make to transform your applications should all be pointed at and informed by shrinking the amount of time that takes you to turn around and realize some business initiative.
That's what we’re seeking with agility. Following pretty closely behind that, you can begin to see why there is a promise in cloud. It saves me a lot of infrastructural headaches. It’s supposed to obviate a lot of the challenges that I have around just standing up the application and getting it ready, let alone having to build the application itself.
So I think that is the view of transformation in terms of agility and why we’re seeing things like cloud. These other things really start to point the direction to greater agility.
... I tend to think that in application transformation in most ways they’re breaking up and distributing that which was previously self-contained and closed.
Whether you're looking at moving to some sort of mainframe processing to distributed processing, from distributed processing to virtualization, whether you are talking about the application team themselves, which now are some combination of in-house, near-shore, offshore, outsourced sort of a distribution of the teams from sort of the single building to all around the world, certainly the architectures themselves from being these sort of monolithic and fairly brittle things that are now sort of services driven things.
You can look at any one of those trends and you can begin to speak about benefits, whether it’s leveraging a better global cost basis or on the architectural side, the fundamental element we’re trying to do is to say, "Let’s move away from a world in which everything is handcrafted."
Let’s get much closer to the assembly-line model, where I have a series of preexisting trustworthy components and I know where they are, I know what they do, and my work now becomes really a matter of assembling those. They can take any variety of shapes on my need because of the components I have created.
We're getting back to this idea of lower cost and increased agility. We can only imagine how certain car manufacturers would be doing, if they were handcrafting every car. We moved to the assembly line for a reason, and software typically has lagged what we see in other engineering disciplines. Here we’re finally going to catch up. We're finally be going to recognize that we can take an assembly line approach in the creation of application, as well, with all the intended benefits.
Evans: ... Once we have done it, once we have removed that handwritten code, that code that is too big for what it needs to be in terms to get the job done. Once we have done it once, it’s out and it’s finished with and then we can start looking at economics that are totally different going forward, where we can actually flip this ratio.
Today, we may spend 80 percent or 90 percent of our IT budget on maintenance, and 10 percent on innovation. What we want to do is flip it. We're not going to flip it in a year or maybe even two, but we have got to take steps. If we don’t start taking steps, it will never go away.