As curator of the SOA Center blog site, Software AG's Miko Matsumura recently renamed his site the Whatever-You-Want-To-Call-It Center in reaction to Anne Thomas Manes declaration that SOA is dead. Hoping to emulate the success formula of the Artist Formerly Known As Prince (who is now known again as Prince), Miko wants to replace the devalued "SOA" moniker with a symbol, perhaps a "W" shape (for "Whatever") similar to the singer/songwriter's unpronounceable symbol [usually spelled out as O(+>)]. In an interview this week with InformationWeek senior editor Roger Smith, Miko said he are also amenable to the acronym AFKAS, which is short for the "Architecture Formerly Known As SOA."Based in Germany, Software AG has almost 40 years of global IT experience and more than 3,600 employees serving customers in 70 countries. The company posted total revenue of $932 million U.S.D. in 2008 by helping more than 4,000 global enterprise customers modernize, integrate and automate their IT systems and processes. As vice president and deputy CTO at Software AG, Miko Matsumura is responsible for the company's technology strategy. Previously, he was VP of SOA product marketing at WebMethods, a company that Software AG acquired for $543 million in cash in 2007.
Matsumura is a service-oriented architecture (SOA) thought leader who recently co-wrote a pragmatic guide for SOA adopters called "SOA Adoption For Dummies" based on his impressive background in real-world SOA engagements. Matsumura holds an MBA from San Francisco State University and a master's degree in neuroscience from Yale University.
InformationWeek: You've written about SOA adoption. Where do you think the industry is in adopting service-oriented architecture?
Matsumura: I'm sure you've seen the Gartner hype cycle, which is the graphic representation of the maturity, adoption, and business application of specific technologies. I think we're on the "Slope of Enlightenment" before we reach the "Plateau of Productivity." Anne Thomas Manes' pronouncement about the death of SOA was made in the depths of the "Trough of Disillusionment." If you look at the timing, it looks like SOA is climbing out of that whole primordial soup. InformationWeek: Where the economy's at obviously didn't help.
Matsumura: I think the economy has forced people to evolve and become cleverer.
InformationWeek: As our recent InformationWeek Analytics Report on SOA detailed, SOA adoption is hard but companies aren't walking away from their investment.
Matsumura: It's timely to ask the questions "Why is it hard?" and "What are we missing?" What our customers are telling us is that they think we're moving toward a more science-driven economy, more quantitative than speculative. This means that we have to grapple with what's there, and this principle of grappling with what's there leads to a more nonviolent approach to cost reduction.
InformationWeek: A lot of IT projects at the moment seem to be driven by the idea of cost reduction. Matsumura: Well, you can either take either the ready-fire-aim or ready-aim-fire approach. The ready-fire-aim approach means you just starting cutting stuff down or shooting things down or cutting divisions. There's a difference between cost cutting and cost reductions. Cost reduction, done intelligently, is based on what people are actually doing, not on what they say they're doing. What people say they're doing diverges from what they're doing. If you can get more quantitative information about what's actually happening, you can do cost cutting in a more scientific manner.
InformationWeek: Which is obviously a more responsible approach to the downturn.
Matsumura: If you're in the spirit to rebuild things, you ought not to shoot first and ask questions later. This last approach is a more violent approach that's commensurate with the level of frustration and disgust. Unless you channel that energy, you can't build your economy on that. You have to pay more attention to what you do and realize your actions will have consequences. There's a ripple effect to the actions you take. Slashing and burning will give you a temporary reduction in your cost but you may also get a permanent reduction in your revenue. You can't blindly chop things away, which some companies have reached the point of doing.
InformationWeek: Our Analytics Report didn't find much slashing and burning of SOA infrastructure.
Matsumura: We've found that people are respecting the investment they're made and are proceeding. That's an interesting thing about SOA. Part of the proposition of SOA is that when you have this mediation layer, then underlying services can be rationalized, redundancy can be reduced, and you can look at your infrastructure and clean it up from a portfolio perspective.
InformationWeek: Silos still need to be traversed ... Matsumura: For example, look at service consumption rates. What I've found is that when I ask a customer, "How many services do you have?," they give me a number. Once we introduce our WebMethods Insight diagnostic tool into the mix, the numbers or services and consumers of services are always underestimated. Surprise factor -- how did people know there was a service available to be consumed? If there is orthodoxy around the governance of on-boarding people to services it's a request-access paradigm. If you're doing that, then you're in good shape. But there are often well-behaved services in an enterprise and services over here that people are finding willy-nilly. Most IT departments didn't even know those were there. For example, SAP is service enabling their ERP app -- which is thousands of services -- and someone who knows what they're doing can access those. There are all these general ledger, accounts receivable, accounts payable services. It's the same for Oracle Financials and PeopleSoft. Ditto for Salesforce or Google or Amazon or any of the cool Internet platforms. Even network appliances like Cisco boxes and routers are starting to come with Web services interfaces. It's a wild scenario and most people aren't aware of all the services that people are bringing into your network.
InformationWeek: It does seem to be the Wild West, which might explain the "shoot first, ask questions later" attitude.
Matsumura: What we see is that cost cutting is often coarse-grained. You can wake up and go to work and find out that a whole division is gone. Unless you know the dependences, it's dangerous to shut things down or kick people out. But this nervousness doesn't always flow up to the executive suite. They look at a spreadsheet or a BI report and say, "This unit is unprofitable, get rid of it. We have to turn in our number for the quarter. They're done."
Operationalizing that decision -- cascading down from that -- becomes pretty stressful and complex. "It's like I haven't used my left hand in a while, let's chop it off." Which explains some of the midlevel panic and urgency.
Executives need to understand that SOA infrastructures have a mediation layer that provides insulation for these decisions.
First, you have the ability to do security mediation between 15 different security concepts that represent 15 different platforms. And it's simple to bind to the mediation layer that will handle the security context.
Second, there's the value of virtualizing a service endpoint. If you hard-bind an endpoint and you unplug the underlying service, then by necessity the wire breaks and the service consumer is out in the cold. This is part of the buffering functions of runtime governance and the mediation layer.
Third, SOA projects are multilayer by nature and they can get derailed because if a service is unavailable people will start pointing fingers at each other. The service is unavailable but the service has enough cross-organizational dependencies that you don't know whose at fault. This definitely harms a project's progress. Runtime governance gives you visibility down through the whole stack trace -- all the way down to the technical service container level. You can also tunnel up through the network and through the interdependent services and you can actually find out which system is performing at what level. This allows you to cut through a lot of the multilevel organizational blame game. You can say, "The slowdown is here!" You can make a better business decision that says, "We need fix that to make it better," or say "That was a harebrained way to do that. We needed to reroute these people to something that will work rather to do something that keeps screwing up."
InformationWeek: As an architectural approach to building systems, SOA is designed to deliver two major categories of value. One is sharing or reuse, and the other is agility or the ability to change more rapidly.
Matsumura: The discover aspect of a SOA runtime allows you a lot of leverage in fomenting a game plan. It gives you the ability to find or kick out a steady stream of fundable projects just by looking at the behaviors in your network, and trying to figure out how to optimize that.
You can use the SOA runtime system to feed continuous information about key performance indicators (KPI) that you've agreed upon across departments and silos in your enterprise. You can also use the runtime system to provide measurements of accountability.
You need, however, to be able to bind the KPI system back into the motivational system -- that is, there needs to be charge-backs or job performance evaluation or bonuses for reuse. You need to take the measurements that are coming out of the systems and bind them to people's incentives.
InformationWeek: From a C-level executive perspective, it seems the best case scenario is to ask questions, get answers to those questions, and keep your guns holstered. Matsumura: It's more a case of measuring twice, and cutting once.
For more information:
We polled readers to assess whether SOA has simply hit an economy-induced bump in the road or is, instead, at a significant crossroads. In this InformationWeek Analytics Report, we drill down to understand the challenges organizations face with their SOAs and assess future directions for those still in the early stages of adoption.