In-Depth
Big Iron Front and Center in Big Blue’s SOA Push
IBM is touting a more abstract kind of workload—the mainframe-as-service-enabled hub
IBM Corp.’s mainframe fortunes are as strong as ever. Big Blue typically cites a number of drivers, not the least of which is a mainframe value proposition that makes it possible for customers to leverage their existing (i.e., “legacy”) assets as they transition to next-generation workloads.
While IBM representatives have always been quick to talk up the importance of zLinux and Big-Iron WebSphere in this regard, they’re increasingly talking up another, altogether more abstract kind of workload—the mainframe-as-service-enabled hub. To listen to IBMers tell it, service-enablement and the service-oriented architecture (SOA) are tailor-made for the mainframe, which has long been under siege from without. The plug-and-pray promise of SOA, which emphasizes flexibility, adaptability, and—above all else—reusability, promises to obviate (or at least mitigate) such a tactical disadvantage.
In other words, says System z GM Jim Stallings, why go to the trouble of ripping-and-replacing your mainframe system when it’s cheaper—and savvier, in the long run—to repurpose it? To put it another way, if you’re going to rip and replace your mainframe, why not do so with another mainframe? The salient point, Stallings and others at IBM claim, is that a service-enabled overhaul of your mainframe applications can give you considerably more bang for your IT buck.
“There definitely are customers who are seeing the value [of the mainframe] for SOA. I met with customers who are using [SOA] to make System z the centerpiece of their information architectures. So much [mission-critical data] is there already, so many of the applications are there already, that if you give [customers] a way to involve these like you do with SOA, they’re going to seize it,” said Stallings, in an interview this summer.
It’s not simply that customers are abandoning other platforms to redeploy their bread-and-butter applications on the mainframe—although that’s happening, too, Stallings insists—but that SOA is one of several forces that can help relieve the siege-like pressure to which the mainframe has been subject for more than a decade. Once the threat from without has been neutralized to a degree, the mainframe can win on its own (long-standing) merits, Stallings maintains.
“The more you load on the mainframe, because of virtualization, and because it can run at utilization rates that are 85 to 90 percent, you get this trifecta on the mainframe. There’s study after study that proves it is the most powerful platform when it comes to cost, efficiency, and performance,” he argues.
SOA: More than Meets the Hype?
SOA seems as if it’s been a long time coming—the SOA hype cycle has been cresting for at least 30 months at this point, and it still hasn’t peaked. Even so, IBM officials say they’re seeing real uptake of SOA in mainframe environments.
Pete Marshall, a Tivoli representative and a market manager in IBM’s strategy group, points to Big Blue’s recent delivery of SOA management tools for mainframe environments as evidence of such demand. “The existing z community [is] not standing still and [is] going out there and implementing on z the kinds of applications that are brand new, and are architected for Web services and so forth,” he says.
“IBM’s focus is that z is one of those great platforms for SOA, opening up access in organizations where the barriers for doing that were just too high in the past, and z is a great platform because it speaks for the availability and the security and the scalability you need for that.”
Enter a new version of IBM’s Tivoli Federated Identity Manager (TFIM) for System z, which Big Blue announced at last month’s SHARE conference. IBM didn’t develop TFIM for System z with the expectation that a market for it would eventually develop, Marshall says; on the contrary, the System z flavor of TFIM was designed to address pressing customer pain points.
“I know this came out of customer demand. We’ve got the initial uptake, and the enterprise customers say, ‘This is great, but what I really need to do is expand from using SOA as an architecture on distributed systems to z, because I’m basically not seeing the whole world if I don’t have my z-based applications in this world.’”In this respect, Marshall says, TFIM helps make SOA safe (so to speak) for security- and availability-obsessed mainframe customers. “The Tivoli announcements that were made at SHARE very much focused on security at that technology stack level. We were focused on adding security on z for SOA applications,” he explains.
“[If you] think of an SOA, you’re proposing multiple services, on multiple platforms, often in multiple networks and sometimes even in multiple organizations, and you need something to manage identities and provide secure access to resources across all of this.”
An SOA Blitz
Big Blue this year kicked up its SOA marketing, starting this spring with the announcement of 11 new products for (and 20 new enhancements to) its WebSphere application server stack, along with an obligatory tie-in via IBM Global Services. IBM’s move came in spite of acknowledgements from executives—such as Steve Mills, senior vice-president and group executive of IBM Software Group—that SOA isn’t quite ready for prime time, at least as far as the requisite application software, development solutions, and management tools are concerned.
That’s not all. Last week, IBM staged another SOA blitz, this time in tandem with academic superstar Georgetown University, which—along with academic fellows George Mason University and the College of Charleston—touted a new SOA training curricula for IT professionals.
The message, IBMers contend, is that even if the SOA technology vision is itself still incubating, adopters should understand that its requisite software, services, and skill components are in the pipeline.
In the meantime, says Big Blue’s Marshall, why shouldn’t companies think about pursuing service-enablement where it most makes sense—that is, in tandem with their hyper-critical mainframe assets. “I think it’s really where the business opportunity [is], and the business opportunity is in combining all of my existing critical applications and recombining them and creating new applications of even greater value, and clearly anyone who has anything on z rapidly comes to the realization that at least some percentage of the critical app and therefore the critical services are running on that platform,” Marshall concludes.
“IBM’s position is ‘Let’s not leave z out of the equation.’ If you’re building new, critical services, we believe you should take a look at z—and certainly don’t leave it as an orphan. There was a tendency back in the client server day to relegate z back around the back end data server. Truth is, there’s a tremendous amount of business logic there, and SOA gives you a great opportunity to use it.”
Related Article: