In-Depth
What’s Really Driving Application Modernization?
IT shops don't always embrace application modernization as a means to ratchet up efficiencies. They're mostly looking to cut costs.
To listen to some industry analysts tell it, organizations are rushing to embrace SOA, Web services, and other application integration or application modernization technologies.
That, however, is probably too optimistic by half. In fact, some organizations don't so much rush as get dragged -- often kicking and screaming -- into modernization projects.
In many shops, IT veterans say, application modernization is a matter of last resort, embraced only when the pain of licensing, maintenance, support, or platform costs becomes too much to bear. In other enterprises, IT shops are modernizing older applications (or swapping them out entirely) to mitigate risks -- such as the pending retirement of programmers, operators, or support personnel associated with their operation.
In both cases, IT shops aren't embracing application modernization chiefly as a means to ratchet up efficiencies, streamline or reconcile business processes, or introduce new (and potentially profitable) services. They're mostly looking to cut costs.
"There are two different paradigms where legacy modernization is not necessarily [done as a means to] adopt SOA, or to [transition] to a new platform. It's done for other reasons, cost being one of them," says Vivek Mehra, vice president of global financial services with Keane Inc., a $1 billion consulting and integration firm based in San Ramon, Calif.
"[Modernization] is mostly [driven] from a cost-reduction perspective, and often from operational risk reasons as well," Mehra continues. In the "legacy" application space -- and the government, insurance, and financial services segments are rife with such applications or platforms -- the latter issue is particularly pressing, he indicates.
"If you have legacy platforms where the code is 40 years old and the knowledge isn't documented, there's a definite risk there. What happens when these [IT pros] retire? Sometimes the answer is simply training more people and transferring [to] several people the knowledge that the incumbents have and making sure that there's a next-generation workforce ready to continue."
Mehra says the decision to modernize an application or to replace it entirely is often based on a sliding scale of costs versus benefits. In other words, he stresses, the ROI case for moving to a new system is relatively simple: the cost of doing so (projected over the course of several years) is cheaper than that of maintaining the status quo. Alternately, there's the incalculable cost of not modernizing and then being left high and dry. Imagine trying to support poorly-documented application code once key knowledge holders retire and you get the idea.
"If you can forecast your earnings and your growth early in the year -- and a lot of these companies can do just that -- there's very little impetus for them to modernize just for reasons of niggling costs. There's got to be a lot of pain there," he comments.
"The one recurring theme is the high level of operational risk, what's going to happen to a system if it's going to be orphaned suddenly. That's a major driver for a lot of these companies."
Mehra notes that the decision to modernize becomes a no-brainer "when they eventually realize that the cost of running these platforms often consumes a significant portion of their IT budget -- so they aren't able to do any discretionary spending. That's often the [impetus]," he says.
The good news for mainframe pros is that Big Iron is no longer viewed as a prohibitively expensive platform, either for ongoing maintenance or support, or as a destination for new workloads. For one thing, Mehra says, organizations have figured out that it's a hugely non-trivial task to achieve mainframe-like reliability and resiliency on Unix, Linux, or Windows platforms.
"I think the mainframe is not in danger, and the reason it's not is because it's difficult for … some of the smaller carriers [and] the mid-tier companies to invest in a robust distributed architecture on a Unix-like platform or Linux or anything else," he comments.
"It's difficult to mimic the scalability and the performance that you get out of the mainframe. As long as you're able to maintain a mainframe and assuming there's no recurring cost for the hardware, then it just becomes a matter of can you afford that, and if you can, then the answer is that you should just continue on the mainframe and build extensions for [the applications you're hosting on] it."
Are there any hard and fast rules for deciding to modernize an application?
"The ROI has to be strong, and there has to be a business buy-in to do anything, because business doesn't want its day-to-day operations to be disrupted because you're converting to a new platform. They don't care how it is enabled, they don't care if it's running on 40-year-old [technology] or new technology, as long as the business gets done," Mehra explains.
"What goes into this decision is all of these factors: it's the business, it's the ROI, it's the presence [or absence] of skill sets in the company, the cost of maintenance."
Companies that modernize their applications simply to realize "efficiencies" or "synergies" or to optimize their business processes might be making a mistake, Mehra indicates. His expertise lies largely in the insurance space, he cautions, but for many of these companies, if they haven't already figured out how to do it faster or better, they'd be out of business.
"There are maybe some cases like that, [where] that would be true for more of the smaller firms. The bigger firms have already figured that out. If their IT is not effective, they would typically be out of business," he explains. "It's the smaller players where they would look to modernize because they want to get additional efficiencies out of their systems."
This isn't to say that big firms can't benefit from business process optimization, Mehra stresses -- just that one of the most oft-cited drivers for application modernization (case study-like performance boosts) isn't all it's cracked up to be.
"That isn't to say that all companies are as efficient as they can be: you can almost always look at a process and get 20 to 30 percent more out of it," he concludes. "There's almost always a better way to do it. The impetus [to do it by modernizing] just isn't always there."