SOA Goes Guerilla
Oct 31, 2007 5:00 AM PT
The latest BriefingsDirect SOA Insights Edition, Vol. 26, provides a roundtable discussion and dissection of service-oriented architecture (SOA)-related news and events with a panel of IT analysts and experts.
In this episode, our group examines the relationship and tension between enterprise-wide SOA and more discrete Web-oriented architecture (WOA) -- what we like to call "guerrilla SOA." We also look at the probable acquisition of Business Objects by SAP, and the recent Information On Demand conference from IBM.
Join noted IT industry analysts and experts Jim Kobielus, principal analyst at Current Analysis; Tony Baer, principal of onStrategies, and JP Morgenthal, CEO of Avorcor. Our discussion is hosted and moderated by me, Dana Gardner.
Listen to the discussion (52:26 minutes).
Here are some excerpts:
On WOA and 'Guerrilla SOA'
JP Morgenthal: You always have two different communities -- one very active, very leading-edge groups like financial services, and then there are always on the lookout for new technologies that are going to help them do their business faster and better, doing lots of more. They are not risk averse they are willing to throw some additional capital at those projects and see what they'll bear. Right now they are the primary community that I see that's really gung-ho on this. ... The other group, let's call them the moderates or the laggards, definitely view this technology as questionable. ... They still lean towards, "I want a complete application. I really don't want to play with this stuff."
Dana Gardner: Therefore, people might be pushing it off, which relates to what IBM is doing, which is trying to make this comprehensive, more simplified, and more integrated, so that, in addition to those cutting edge organizations in such fields as financial services, this could be more palatable for the larger bell curve of enterprises.
Morgenthal: My concern for the industry as a whole is that people are going to view it as throwing a lot of consulting dollars down the drain and not seeing any value for it. I've recently joined the camp, at least academically, not in any way physically or throwing my weight behind it, but Guerrilla SOA is what I have been doing in my business. I just haven't put a title to it.
Gardner: A year or year and a half ago, we started seeing WOA ... and it's also been called Enterprise Web 2.0 or Enterprise 3.0. But, it's really putting emphasis on REST, as a way of leveraging HTTP as a Web service. And now WOA is becoming more of an emerging best practice. Guerrilla SOA better captures what it's up to or about than WOA. ... So, this notion of an application with a REST style for building Web services based on straight HTTP and XML sort of applies to what JP has been talking about.
Morgenthal: I'm much more in favor of small, non-enterprise oriented, focused projects that deliver value within 30 to 90 days. I see that's the greatest value right now for using these technologies based on SOA, Web services, and the like, because the enterprise stuff is nice, but right now it is too fluid for the industry to grab hold of. It's resulting in potential large-scale problems for companies that have no idea how to build the distribution. ... The problem with distributed environments is that very few people actually know how to manage them.
In IBM's case, they are one of the founders of distributed computing. At their core, they understand it well, but they buy too much into their own marketing hype and don't tell customers well enough, "Hey, look, at the core of all this, of what you're trying to do, trying to get more agile, we lived there. We built the first computers that became agile and communicated across network."
Tony Baer: Parts of this argument we could have had 10 years ago, the whole idea of the big umbrella vendor. If nobody wanted a big umbrella vendor and wanted best of breed, SAP would not be what it is today. I remember during the emergence of the ERP (enterprise resource planning) market about 10 or 12 years ago, there was a debate: "Shall we go best of breed, versus an umbrella approach?" The market has clearly spoken.
However, what you've gotten at the same time is a revolution that picked up steam with the original Borland IDEs and the popularity of bottom-up development, and was energized by the original Visual Basic. There is a powerful constituency of organizations that need guerilla SOA and need to get it done now. It's also behind the rise of agile development.
When you look at just the difference in style between conventional Web service and RESTful, there is a little bit of an irony. Conventional Web services were touted as a simpler alternative to an earlier incarnation of SOA, which was CORBA. This reflects a growing maturity in the field. As we started getting a little more experience working with some of those Web-services technology, we realized that maybe we didn't always need those complicated SOAP headers. So, why not dispense with that, because most of our needs right now are for simple things like fetching data.
Jim Kobielus: Everybody is going to run around the corporate standard, if the corporate standard doesn't meet their needs. It's the actual knowledge workers, the end users. If IT can't give them what they need, they are going to find it some other way. If what the knowledge worker needs is not being funded out of capital budgets and being supported by IT, they're going to pay for it our of their monthly expenses. They are just going to grab it for free on the Internet and mash it up.
Morgenthal: I've been focusing more on the small and mid-sized market, and these guys just want to get something done. The interesting thing is that they don't spend their time sitting there wondering, whether they're going to do Web services or SOA. It's more like 1,500 calls coming in a day, they're being bombarded, and yet they still have to get stuff done. So, it's the backlog.
Then you come in and you tell them, "Hey, in three weeks I can give you a completely new wrapper around everything you have, leave exactly what you have in place, but allow you to do everything you wanted to, the way you want to do it." At first, they say, "Right, show me." Once you show them, it opens up a non-stop flow. They get it the minute they see it.
Gardner:The Web itself and the WOA and the guerrilla SOA are all part of the same trend, which is away from the need for a large corporate IT umbrella, but that you can get things done, satisfy customer needs, be innovative and agile in new markets, and can go global -- all based on not needing one big umbrella, but leveraging what's able across a rich, fertile, open ecology.
On IBM's Information On Demand Conference
Kobielus: About two years ago, IBM established an organizing framework for their data management, database integration, and other data solutions, called "Information On Demand," and it's just a big catch-all for the products they already had, as well as lots of new projects they've been developing to address data management under the SOA big top.
IBM released 10 press releases, and even those press releases didn't capture every nuance of every product announcement, enhancement, and initiative they've got going on. ... They announced upgrades or enhancements to their databases, data warehousing products, their master data management (MDM) products, their data integration products under the Information Server portfolio, their enterprise content management products, the FileNet products, plus the preexisting IBM content management products.
Baer: What's really interesting is the whole idea of IBM biting off more than they can chew. IBM and Oracle are among the few organizations that could pull off something like this and not be overwhelmed by it. You and I saw this several years ago when Ascential had it's analyst conference right after the acquisition by IBM and they revealed the roadmap. What's impressed me is that it's been a very deliberate plan.
A cornerstone of that was Information Server, the whole information-server strategy. Ascential itself was kind of a mini IBM, a company that was glued together by acquisition. What they realized was they had all these disparate tools that ultimately related to the lifecycle of data in all those different forms, and, prior to the acquisition by IBM, they had a roadmap which, I believe, was called Hawk and Serrrano.
The interesting part was that it was all going to become metadata driven and that would drive all the data-integration and data-access strategies. So, I see that as the unsung hero of all this. It provided a more global perspective IBM needed and rationalizes all of these other initiatives. It's not that everything is acting off of Information Server as a hub, but it provides a logical core or gut unification theory.
On Business Objects and SAP
Kobielus: The big news there of course was the pending acquisition by SAP. One of the good things was, at the very start of the keynote, Bernard Liautaud, the founder of the Business Objects, reassured the customers, employees, and partners that Business Objects will be a standalone product group under SAP. It will autonomous. It can continue to pursue its vision. ... It will up to Business Objects whether it makes sense to use a particular piece of SAP technology in any given product, but he reassured everybody that there will be growing integration between Business Objects and SAP offerings.
But [Henning Kagermann, chairman of SAP] intends to have it both ways, because he then said, "We will also make sure that Business Objects maintains an equivalent level of tight integration with all of our competitors." He's trying to have it both ways, but at least Kagermann was speaking the right speak. From my discussions afterwards, everybody said, "Yeah, I think they are speaking in good faith. So far, so good. We'll wait and see." The deal has not been closed yet, and it will be a couple of months.
I see the convergences that are going on are all being driven by SOA mega brands that are continuing to bulk up on the full range of best-of-breed tools that enterprises are asking for. SAP, although it has BI, data warehousing, and data integration under NetWeaver, none of that is best of breed. It's all primarily just in the box when you license their CRM or ERP applications.
Dana Gardner is president and principal analyst at Interarbor Solutions, which tracks trends, delivers forecasts and interprets the competitive landscape of enterprise applications and software infrastructure markets for clients. He also produces BriefingsDirect sponsored podcasts.