Wednesday, November 15, 2006

Transcript of Dana Gardner's BriefingsDirect SOA Insights Edition Vol. 4 Podcast on Oracle OpenWorld and Oracle's Linux and SOA Support

Edited transcript of weekly BriefingsDirect[TM] SOA Insights Edition, recorded Oct. 27, 2006.

Listen to the podcast here. If you'd like to learn more about BriefingsDirect B2B informational podcasts or to become a sponsor of this or other B2B podcasts, contact Dana Gardner at 603-528-2435.

Dana Gardner: Hello, and welcome to the latest Briefings Direct SOA Insights Edition, a weekly discussion and dissection of services oriented architecture (SOA)-related news and industry events with a panel of independent IT industry analysts. I’m your host and moderator, Dana Gardner, principal analyst at Interarbor Solutions.

Our panel this week consists of a return by Neil Macehiter. He is research director at Macehiter Ward-Dutton based in the U.K. Also returning is Steve Garone, a former program vice president at IDC, and a founder of The AlignIT Group, and now an independent industry analyst. And joining us for the first time, we’d like to welcome Joe McKendrick. He is an independent research consultant, and a contributing editor and columnist for Database Trends, as well as a SOA blogger for ZDNet and ebizQ. Welcome, Joe.

Joe McKendrick: Thanks, Dana.

Gardner: We’re going to talk this week -- which is the week of Oct. 23, 2006 -- primarily about Oracle OpenWorld. This large trade show, focused on Oracle’s ecology and products, was held this week in San Francisco, and some interesting news came out of this event, primarily around SOA, development, middleware, and even Linux. So let’s pop over first to Joe in metro Philadelphia. Joe, tell us what you saw as interesting or impactful from the Oracle OpenWorld Show, particularly vis-à-vis SOA?

McKendrick: Sure, Dana. Well, it was definitely a big show, and they really made a big splash. Just let me say, Oracle has been confusing the heck out of the market over the past year or so with its double Fusion initiatives. There’s the Fusion Middleware, of course, and then there’s the fusion of Oracle’s acquired companies, such as PeopleSoft and Siebel. Are the two fusions connected? Who knows?

I do research for two of the leading Oracle user groups, the Independent Oracle Users Group, IOUG and OAUG, The Oracle Application Users Group, and I find few end-user customers know what to make of Fusion, or how it’s going to impact them. I’m not quite sure if OracleWorld brought total clarity. I think it might have brought some clarity to the question. What we want to talk about here, of course, is Oracle’s Fusion middleware, which I equate to IBM’s WebSphere on some levels.

In a survey I conducted and put together for IOUG earlier in the year, only a small handful of Oracle database administrators and users could see how Oracle Fusion middleware could help them, or lower their deployment costs. So Oracle has a lot of user education to do, and it looks like this week’s OracleWorld was a good start, but there’s plenty of work ahead.

Gardner: It seem to me that Oracle at this event really put a stake in the ground and said, “We’re not just a database company, and we’re not just a database company plus business apps.” I think they were really saying, “We’re going to be a full-service software company from soup to nuts, and development, deployment, and lifecycle.” So, that really puts them up into the IBM arena. Would you agree with that?

McKendrick: I agree with that, and it's interesting, because looking at what’s been coming out of OpenWorld, you not only have IBM, but you have two of the largest vendors in the space -- Oracle and Microsoft -- having made recent announcements that are diametrically opposed in their proposed approaches to SOA. Microsoft had their own SOA confab a couple of weeks back, on the one side saying, “Go slow, build incrementally, and don’t spend a lot of money on your SOA as you roll it out.”

Then at the other end, you have Oracle, which is all about big SOA, saying “Take your huge, 10-terabyte database, take your huge ERP system, and move some of that functionality out to a service layer, preferably Oracle’s.” Plus, Oracle was talking a lot about mainframes, recognizing that IBM has been quite successful in reenergizing this big server platform, and Oracle wants a piece of that action as well -- more big SOA.

Gardner: I also did some field research recently that came up with a pretty solid number. It’s a general number, but of the SOA services that are in production and in use that we could get some insight on, fully two-thirds of them were of a services data layer nature. That is to say, they were all about making data available from a variety of sources, repositories, and applications to then be used more generally. Therefore, the whole role of data seems to be that the horse sits in front of the cart when it comes to using SOA, which of course would put Oracle in a pretty good position.

McKendrick: Absolutely. They’re also strong in applications as well, but it's interesting, Dana, as you point out, that there’s a convergence in place between the world of enterprise data management, and the world of application management or SOA. The two have been in separate silos, so to speak, with different disciplines, different types of professionals working on each, but inevitably they’re going to come together, and Oracle sees that on the horizon.

Gardner: Let’s go next to Neil, tell us what you took away from the announcements emanating from Oracle OpenWorld?

Neil Macehiter: A couple of things. First, back to your comment around the two-thirds of the SOA initiatives focusing on the data layer. I must admit that doesn’t surprise me, because the proliferation of information, and the variety of information sources that organizations have, and the challenges they’re facing around things like compliance and getting better insight into business operations absolutely depend on getting access to information.

I think that a SOA approach to that is something that can help them address those challenges. And interestingly, given where Oracle comes from, and its strength in the database; it’s actually an aspect of the service oriented strategy that they haven’t talked up significantly if you compare it, for example, to IBM around its whole information server strategy. That would be my observation there.

In terms of the broader recent announcements of Oracle, clearly they’re upping the ante in terms of their credibility in the SOA space, with one eye on IBM, another on Microsoft, and their third eye -- or their eye in the back of their head -- on SAP with NetWeaver. It was interesting, looking at the announcements they made particularly around the SOA suite, which is their bundling of their BPL Process Manager, Business Activity Monitoring, JDeveloper, Business Rules Engine, where they were really talking up the governance aspects of their offering there.

This is something that historically Oracle has not spoken greatly about. They didn’t make a big push there, and that is no surprise given that governance has been such a hot issue recently. I was also interested in the announcement around their WebCenter, which is the coming together of their portal capabilities with more Web 2.0 style technologies around RSS, Wikis, etc. This is still in its comparatively early phases, and is primarily oriented toward the application developer community, in terms of being able to build these mashup style applications that merge content ... to enterprise applications, and support what some people refer to as situational processes. That’s something that IBM has been talking a lot about around its QED Wiki technologies.

Gardner: What about Joe’s point that Fusion is been sort of fuzzy? Now, some people were making a hypothesis that the reason that Fusion was fuzzy, or confusing, was because Oracle was going to make some acquisition. Perhaps, they'd acquire JBoss, perhaps BEA, and bring in this capability for their middleware and development and deployment power from an outside perspective. I guess that’s still possible. Did you get a better sense -- and you can pipe in here too, Joe -- that they’re defining Fusion as the end-all, be-all for their middleware and development?

Macehiter: Well, I think Oracle is clearly positioning the Fusion middleware as their core SOA infrastructure offering, to allow organizations that don’t have Oracle applications to build services-oriented style solutions -- or if they do have Oracle applications to extend those in the same way that SAP, for example, is pushing NetWeaver. Fusion middleware is an enabler of Fusion’s application convergence.

My own view is that Oracle, with its notion of "hot-pluggable" -- and looking at the way it’s made acquisitions, and to my mind some pretty smart acquisitions, around things like, security and identity management, and Web services management -- this is the core platform. They will look either for tactical acquisitions to plug gaps, for example, in the areas around governance or extending some of the management capabilities, but I can’t see them going beyond that. The alternative will be this "hot-pluggable" architecture, which is an attempt to allow them to hook into other capabilities, be they open-source type offerings or solutions from other vendors.

Fusion is the way forward for Oracle now, and that’s where they’re investing their effort. They need that as well, given that it’s going to be the foundation of the unification of Siebel, PeopleSoft, JD Edwards, and Oracle E-Business Suite.

Steve Garone: If I can just jump in for a moment to add a couple of short comments. What I just heard makes perfect sense. I may be a bit of a contrarian here, but I’ve never actually found Fusion, either as a strategy or a product set, very confusing. It was very clear to me that this was Oracle’s way forward in terms of infrastructure and applications. The only confusing part of it in the past has been, a sort of a late incoming message from Oracle that their middleware is going to be “hot-pluggable,” meaning you can come in with elements, tools, and so forth, from other sources, plug them in, and they will work -- and everything will be integrated in a nice fashion.

That was not an approach that Oracle talked about three or four years ago and, in fact, they sent out the opposite message. This is really where the confusion lies. They pretty much cleared that up. The only issue is, do their actions speak as loud as their words do, and is in fact the Fusion middleware stack as “hot-pluggable” as they say it is.

Gardner: They did certainly embrace Eclipse and moved toward open source tooling.

Garone: Begrudgingly at the beginning, I believe.

Gardner: Well, it’s against their nature, I suppose, but everyone has had to do that. It sets us up for this Unbreakable Linux announcement -- a bit of a shocker toward the end of the week-- where [Oracle Chairman and CEO] Larry Ellison got up there and said that they’re going to create their own distribution equivalent to the Red Hat Enterprise Linux distribution. He said they’re going to support it for a lot less money then you might have to pay Red Hat, and they’re going to tell you that it's configured and it’s indemnified, and they’re going to back that up.

So, that puts Oracle in a much larger position as a platform support vendor. It changes the business model to one of subscription and maintenance support, and puts them in a position to start being a bigger partner now, for a variety of different configurations of stack, be it partially open source, partially Oracle -- or for other partners. It seems to me that they’ve really taken a big step toward a much larger software provider role. Any reactions to that, Joe?

McKendrick: Well, Oracle’s been trying to get into this operating system space for a long time. In the past decade, they were one of the aggressive promoters of the whole network computing concept, where theoretically you wouldn’t be relying on a fat client, so to speak, meaning Microsoft Windows. You could access all your applications via some type of thin client or network computing device. They very aggressively were at the forefront of those efforts, as well as other servers, I guess you could call them, preformatted packaged database servers, where you wouldn’t need to worry about the operating system. This is an interesting move, and an interesting end-run around a major Linux provider.

Also, what I find interesting from OpenWorld is their agreement with IBM to support implementations of the Oracle database on an IBM mainframe. Now, this is very historic, in a sense.

Gardner: Yeah, co-opetion in the best sense, right?

McKendrick: Exactly, no one in the mainframe world could have imagined that someday Oracle would be running on the mainframe. It’ll be running within a Linux partition in the z/OS Operating System, or rather the Linux portion of the zSeries.

Gardner: I think the logic there for IBM is, “Listen, if we can’t get DB2 in the account, at least we’re going to get the hardware, and we’re going to get people locked into the zSeries for a long time. We’re going to make some good maintenance money, because, heaven knows, once you get a large, honking relational database in a mainframe it sits there for a long period of time.” I can see the logic of it from both sides, but it is interesting that they’d actually extend hands across the table, and make it happen.

Macehiter: I think it will be interesting if you look internally within IBM to see from where that relationship with Oracle was driven, and my guess would be that it was primarily driven by global services. They’ve been working incredibly hard to increase the usage of mainframe MIPS, by partnering with a variety of vendors that are in direct competition with IBM.

Gardner: Probably not out of the DB2 group, right? Do you think we can read in the tea leaves here -- a little bit by what Oracle has done and by what IBM is doing as well -- that, the heyday of the relational database as a core lock-in, and as a high-margin, expensive aspect of a data center and an IT architecture is coming to a close?

Is Oracle saying, “Yeah, we recognized that lock-in doesn’t work, and so we need to provide the low-cost, best-configured approach, and our real new value proposition is in our business applications, and our ability for people to create data-driven applications quickly, easily, and cheaply?” And, does that mean that at some point in the not too distant future that the cost for relational database, be it an Oracle, a Sybase, an IBM DB2, and -- because of the market forces -- the Microsoft SQL Server, are going to plummet?

And, that perhaps an open-source subscription model for databases is in the offing, particularly as we get into virtualization and multi-core, which makes it very difficult to charge for these things on an upfront license basis to begin with. So, let’s go around the table here. Are we at the end of the heyday of the relational database?

Macehiter: Well, I spent six and a half years of my career working for Oracle, and two and a half at Sybase, in the heyday when the database was absolutely the control point. If you look at what’s happened in the market, the industry has consolidated massively.

Sybase is now really more in the mobile space than it is in the core database space. Ingres is now open source. You’ve really only got Oracle, IBM, and Microsoft with SQL Server vying for dominance in terms of the commercial database. So, you’ve had significant consolidation. You’ve got the entry of, the likes of, MySQL, which now is offering much more of an enterprise offering with its support network.

I think all the signs are there that it’s becoming commodity technology, but that should naturally result in a driving down of cost. I think the organizations also depend more and more on unstructured information that doesn’t exist inside of a relational database. So, everything is there that points to there being a drive down in cost, and it's interesting in that regard, and also in regard to the Oracle Unbreakable Linux, where Oracle makes the vast majority of its profit in its maintenance services. The control point is shifting, and it’s much more about how you get the information out and harness it in the business processes and activities that support employees, customers, and partners in their operations. I think absolutely the writing is on the wall.

Gardner: So, it’s not so much about the object model. It’s about the business model, right?

Macehiter: Indeed.

Garone: Looking forward, we’re going to see a lot of emphasis put on Fusion by Oracle, which really brings together the leverage point of its product strategy, which is applications and middleware. I agree with everything Neil said, I think it’s important to know that what’s going on here, and to some extent it's mimicking what is going on with a lot of other software-type products or markets that have seen open-source offerings come in and have their effect on what you can charge, and how you license, and so on.

So, I don’t think it’s unusual in that regard. It’s a little bit difficult for me to describe database technologies in companies like Oracle as commodities. I don’t really see them that way, because “commodities” implies there’s really not a lot of differentiation and value-add. You’re just kind of getting the same thing from everybody at a bare-bones price. I don’t really think that reflects the database market today, or even the database market down the road. I do think that there are, and will continue to be, other competitive forces that are going to drive Oracle to look elsewhere for its high profits.

Macehiter: I would. Looking at Microsoft’s recently announced quarterly results, if you look at the grades of SQL server from Microsoft, those numbers were phenomenal, again.

Garone: Look at what happened with Oracle and IBM’s recent financials as well. They’re making a ton of money, and it’s a high growth, in the very area that we’re talking about being commoditized. This hasn’t happened yet, but I certainly see these companies setting themselves up for that.

Gardner: What do you see, Joe?

McKendrick: I think Neil and Steve have definitely hit on two important things -- essentially a double-whammy -- that’s impacting the industries. The open-source databases, in particular open-source middleware, are coming into play. Fusion, the whole SOA phenomena, is still getting off the ground, but I think there’s some long-term implications. The companies are making money off of these products, but I think they’re looking long-term down the road.

If you look at what’s happening in both the database and the ERP world, Oracle’s bread is databases, and its butter is ERP applications as E-Business Suite. Many of the functions of the database are, or will be, abstracted out to a service layer, and eventually the same thing may happen with ERP systems, where many of those functions get abstracted to some type of what we call the SOA layer. Granted, it’s going to take a few years -- we’re probably talking about five or ten years down the road -- but these are large ships, and they take a long time to turn around, and they have to look at what’s happening in the space.

Oracle rightly recognizes that the future lies somewhere more out in the middle, not in the big complicated mega-databases, and mega-ERP systems. Recently, in my ZDNet blog I quoted Bruce Richardson of AMR Research, who predicted that, within a few years, probably by 2010, SAP and Oracle customers may actually stop buying applications from their ERP vendors. Some folks have said Bruce may have gone a little too far with his predictions, but I think there’s some credence to it. I’ve actually spoken to customers who have been able to deploy SOA techniques as workarounds to more expensive, painful upgrades to their ERP systems.

Gardner: We’ve already had a little bit of a discussion around this, and I refer our listeners back to an earlier Briefing Direct SOA Insights Edition Vol. 2 for the discussion about SOA and ERP, and how they shake out.

Returning to OpenWorld, one of the things that jumped out at me too was this new warm, fuzzy Oracle: “Oracle’s your buddy.“ That wasn’t necessarily the case in many peoples' minds. We saw [Sun Microsystems CEO] Jonathan Schwartz, [AMD Chairman and CEO] Hector Ruiz, and [Dell Chairman and CEO] Michael Dell on stage. We had HP [Chairman and CEO] Mark Hurd on stage -- so a lot more of this big ecology, fuzzy, warm thing.

Also, quite a bit about Web 2.0 as a catalyst in Oracle’s Fusion. I guess when we talk about Web 2.0, we’re thinking about the use of rich media, social media, the links, and viral approach to reaching, and communicating and collaborating, including RSS, Wikis, blogging, and the whole idea of open rich platforms and suites and features to support that. And, that becomes an element of development, and even becomes a development of how services can be consumed and used.

I’ve heard people say that we really shouldn’t have this SOA repository-register approach, that we should have a wiki, and that people will actually go in and be able to, as business analysts with proper rights, be able to decide, and have input to how services will be used and consumed. And that to try to make services governance it too control-and-command and rigid is the wrong way -- that we should make it open and fluid. Like you would with the definition of an encyclopedia entry, as in Wikipedia, you should be able to do that with the definitions and usage of a service.

Any reaction out there to this one, warm, fuzzy community-approach from Oracle, and then secondarily the role of Web 2.0 capabilities?

Macehiter: I think about the Web 2.0 thing in couple of ways. One, is that it was interesting to see that Oracle has moved away from talking about SOA 2.0 which is very fortunate indeed. In terms of Web 2.0 style technologies and approach and their impact on the enterprise, as with many things, this is going to be a mix of the very structured, highly controlled, registry-repository governed approach to service orientation, where it's needed.

For example, in some processes, subject to regulation, those processes need to be highly structured and highly monitored. Equally, what we’re now seeing is really a recognition that organizations work in fluid ways as well. They are highly collaborative, and involved in a lot of interactions, much of which has gone on in organizations in the past, but out of the purview of IT, because IT hasn’t been able really to provide the support through systems for the way that people really work.

I think that’s where the role of these Web 2.0 style technologies can have a role to play in actually enabling IT to better support the way that people often work, particularly dealing with exceptions to these well-structured processes. I think the challenge there is about culture, and getting the frame of mind and reference within the users, so that they can actually utilize this stuff, and not be afraid to use these capabilities.

So I think it's going to be a mix of both, and what we’re seeing is -- with the Oracle’s announcement of this WebCenter technology, what IBM is trying to do around the QED Wiki, and what Microsoft’s doing around things like SharePoint and the Atlas Framework -- you’ve got the dominant providers of IT for the enterprise actually mixing these two worlds around common platforms, be it .NET, Fusion, or WebSphere in IBM’s SOA foundation. So, I think it's going to be a mix-and-match of these two capabilities.

Gardner: What do you think, Steve? Have we got oil and water here when it comes to SOA and Web 2.0? How do you see them relating?

Garone: I agree with what Neil says, and some of it's too early to call. But I think there is a lot of synergy between the SOA approach and Web 2.0. I think there will be an effort to keep things as simple and low-complexity as possible. I see a lot of synergy there, and we’ll see how it goes down the road. It’s just really early in the adoption to putting stakes on the ground at this point.

Gardner: And how about you, Joe, as a blogger? Do you think this is a fad or here to stay?

McKendrick: It’s going to percolate up. When I think of SOA and Web 2.0, I think of that Apple commercial. You see the two guys, one guy’s an up-tight, straight-laced guy in a suit…

Gardner: "I’m a PC," yeah.

McKendrick: And the other guy is the cool Apple dude, hip and laid back. The guy in the suit could be SOA, and the cool hip guy, Apple guy, could be Web 2.0. And they’re trying to understand each other, at least the SOA guy is trying to understand what the Web 2.0 guy is up to, and why is it that he’s so cool and hip, and can’t he be more like him.

Gardner: So, maybe if they were in therapy together, they might actually work together. It might be kind of cool.

McKendrick: Exactly.

Macehiter: I’m sitting here in front of a Dell laptop, and a Mac Mini, so I’m obviously a bit schizophrenic.

Garone: That’s a really interesting analogy though. The commercial that you just described does point out that the two heads have to get together somehow, and see how they’re going to make that shake out, but I think one’s going to support the other. I really do.

Gardner: And, talk about commoditization, I should think that as companies like Oracle and others latch on to these capabilities that the whole idea of blogging as a function, and having an RSS capability as a function, is just going to be immediately embedded within the background here. But it’s the power of the people and the process that rises above it, and so there’s no need for lock-in, there’s no discussion about who has got the better word processor for blogging. It’s just going to be accepted and embedded into these processes up and down the whole development and deployment lifecycle.

Garone: Embedded as services.

Gardner: As services -- you don’t even have to own this stuff yourself. You can just use Blogger off of Google servers if you want.

Macehiter: Absolutely. It comes down to the sort of environment that you can provide, which is where the likes of IBM, Microsoft, and Oracle are trying to drive things, to make it easy to compose and exploit those capabilities in a way that reflects the pace of change, in terms of the way these activities occur within organizations, and how employees really work. You can’t afford to now throw those requirements over to IT, wait six months, and get the solution back.

The individuals need to be empowered to make the changes, and that’s where Oracle, for example, is trying to drive things around this WebCenter, and the competition framework they’re talking about. IBM has similar capabilities around its QED initiatives. It’s around the competition environment, and you’re not going to compete on whether you’ve got RSS support or not, or it’s your RSS, instead of anyone else. That doesn’t make any sense.

Gardner: As someone mentioned earlier, Oracle’s chief competitors are IBM, Microsoft, and SAP, and in doing these things, and with their Linux move, they take the oxygen out of Red Hat. Red Hat’s ability to price with impunity in the market, which it’s had for a bit now, is quickly eroding.

Novell came in and probably tried to do something about it, but didn’t have sufficient clout, at least in the North American market, to do that. Oracle could very well have an impact on the pricing of a Linux support license. With this Web 2.0 stuff, they really are taking a bit out of Microsoft’s quiver, in terms of their set of features and functions.

If you don’t need Microsoft products installed for doing Web 2.0 and SOA then what do you need them for?

Macehiter: Absolutely! That’s the way it’s driving. Interestingly, in terms of the Red Hat thing, it also does raise the question of what IBM is going to do now. How is IBM going to react to that, given that it provides a fair amount of support around Red Hat?

Gardner: Good point.

Garone: From what I’ve read, a good part of Oracle’s rationale for doing this is they don’t talk about it in terms of, “We’ll give you a better price than Red Hat can, but we will do things for you, if you are in fact an enterprise user. You have special needs in terms of upgrades and getting things fixed quickly and easily, and so on, that Red Hat isn't going to do for you, but we will.” So, they’re kind of positioning this as, “Red Hat’s not enterprising -- we are.” I think that’s interesting in terms of your question about IBM, because IBM, obviously, can make the same claim as well.

Gardner: Do you think we can perhaps look in hindsight, with Red Hat buying JBoss, that that was a mistake, that it allowed them to then be viewed as a competitor rather than, people like IBM and Oracle saying, “Okay, we’ll live with that. We’ll let them control that domain. We’ll cede onto them this Linux licensing thing.” But as soon as they got JBoss, then, boom, they got a big red bull’s eye on their back, right?

Macehiter: I think it was either the day after, or a couple of days after, the Red Hat JBoss announcement that Larry Ellison was interviewed in the Financial Times, and it became pretty clear that he called out what has now ultimately happened. He was making comments about once open source gets good enough, Oracle will just take hold of it, and they did that with Apache. Oracle had its own web server embedded within the Oracle application server, and then as soon as Apache became good enough, he just ripped out his own Web Server and put in Apache.

I think Ellison flagged this and I don’t think it’s coincidental that he flagged it about the time that Red Hat acquired JBoss. It changed the rules in terms of the relationship Oracle had with Red Hat.

Gardner: And IBM has done the same with tools, vis-à-vis Eclipse, and then also in buying Geronimo, going with Glucode, rather coming out with a lower-end approach to open source. I think IBM has hedged its bets very well, and can either turn up the spigot, or turn it off, in terms of which direction and which balance to go open source in. Of course, they have a hardware business to fall back on -- which Oracle does not, nor does Microsoft -- and they have professional services revenue. So, I think they’re in a pretty good position to manage this pretty well.

Garone: Just to declare though, IBM’s acquisition of Geronimo took place significantly before the JBoss acquisition. So I know they weren’t responding to that, unless they suspected it happening down the road.

Gardner: No, I think they were responding to the fact that they need to have both an open source and a commercial strategy.

Garone: Correct. It’s a good strategy, and the only gotcha there -- which IBM has been very clear it’s committed to doing -- is making sure that compatibility exists up and down the WebSphere line, which includes the WebSphere products that are based on Geronimo. I’m not convinced that has happened yet, but IBM has stated a very strong commitment to it, and they’re usually pretty good about fulfilling those commitments.

Gardner: As we’ve discussed in the past, these vendors have been able to follow a certain maturity of commercial products that then has a following counterpart in open source. What’s unique and interesting about SOA is that we’re getting not just a following of a commercial process, but we’re getting an open-source process, almost from the get-go.

So how these companies react to SOA componentry as open source is still an interesting conundrum, I suppose. And we’ve discussed that in the past. Joe, you haven’t had a chance to chime in on this. Do you have any thoughts -- and we’ll make this sort of our closing topic: open source and SOA vis-à-vis what we’re seeing from these large vendors.

McKendrick: Some good points raised here. I think middleware is becoming the battleground. We talked about the abstraction of database functionality and ERP, large application functionality into the service layer. And this is where the battleground is going to be in.

Oracle sees this with JBoss as a huge threat. Oracle talks about making its Fusion components “hot-pluggable” with other software from other vendors, but, make no mistake about it, Oracle wants Oracle customers doing SOA the Oracle way. Oracle has this huge built-in customer base, and it has expanded it greatly over the past year, and its likely many will eventually move to Fusion. But Oracle really has to make a good case as to why customers should consider sticking with Oracle up and down through the stack, and not go on with a commodity or an open-source middleware type of solution. That’s a big challenge all the large infrastructure vendors face.

Gardner: Well, there it is. I would like to thank everyone for joining us. We forgot to do this last week, and that was because we were running out of time, and I was afraid our hour-long conference call was going to end and cut us off. What I would like to do at this point is to go around the table and ask everyone to give us a disclosure. Tell us what companies you’ve been working with, so that we are on the up and up here. Neil, why don’t you start?

Macehiter: Yeah, sure. In terms of the vendors that I’ve mentioned today, Microsoft, IBM are customers, or have been customers of Macehiter Ward-Dutton.

Gardner: And how about you, Steve?

Garone: IBM, Oracle and Microsoft.

Gardner: And Joe.

McKendrick: IBM and Oracle.

Gardner: And for me it’s going to be Eclipse, HP, and IONA, and that’s it for this particular show.

All right, thanks everyone. Also for our listeners, if you’re interested in learning more about our BriefingsDirect business-to-business informational podcasts, or this particular show, the SOA Insight’s Edition, to become a sponsor or to learn more about it feel free to call me at 603-528-2435. Thank you. This is your producer, host, and moderator, Dana Gardner, principal analyst at Interarbor Solutions. Thanks to our guests, and come back and listen again next week.

Listen to the podcast here.

Transcript of Dana Gardner’s BriefingsDirect SOA Insights Edition, Vol. 4. Copyright Interarbor Solutions, LLC, 2005-2006. All rights reserved.

Sunday, November 05, 2006

Transcript of Dana Gardner's BriefingsDirect SOA Insights Edition Vol. 3 Podcast with Analysts Steve Garone, Neil Ward-Dutton and Guest Jeff Pendleton

Edited transcript of weekly BriefingsDirect[TM] SOA Insights Edition, recorded Oct. 20, 2006.

Listen to the podcast here. If you'd like to learn more about BriefingsDirect B2B informational podcasts or to become a sponsor of this or other B2B podcasts, contact Dana Gardner at 603-528-2435.

Dana Gardner: Hello and welcome to the latest Briefings Direct SOA Insights Edition, a weekly discussion and dissection of services oriented architecture- (SOA)-related news and events, with a panel of independent IT Industry analysts and guests. I’m your host and moderator, Dana Gardner, principal analyst at Interarbor Solutions. This week, the week of October 16, 2006, our SOA panel includes Steve Garone. Steve is an independent industry analyst, a founder of the AlignIT Group, and a former vice president at IDC. Welcome again to the show, Steve.

Steve Garone: Thanks, Dana. It's great to be here, and I’m looking forward to some spirited discussion.

Gardner: It’s always spirited on SOA, and no one agrees on anything, right? Also joining us for the first time -- and welcome -- is Neil Ward-Dutton, a research director at Macehiter Ward-Dutton, an IT analysis and consulting firm based in the U.K. Hi, Neil.

Neil Ward-Dutton: Hello, Dana, thanks for having me on.

Gardner: My pleasure. Also joining us as a guest is Jeff Pendleton, a former IT executive who has filled many roles, at BEA, HP and other firms. Welcome to the show, Jeff.

Jeff Pendleton: Thank you, Dana. Pleasure to be here.

Gardner: Thanks. Give us a brief rundown of your background, Jeff, so people can appreciate where you’re coming from.

Pendleton: Well, I’m pretty much a jack-of-all-trades. I spent 16 years at HP, alternating between IT, marketing, and operations management. So I got a pretty good, 360-degree view of HP, and how it approaches the world. I made a run at a startup, spending a year in the garage, working on a SOA appliance startup with a bunch of really smart guys from Cisco, just before Cisco announced their whole Application-Oriented Networking (AON) initiative.

So I played around with that. I picked absolutely the wrong time to try to launch a startup, in 2000-2001. Then I moved over to BEA about four and a half years ago and ran their IT group, ran their education services division. And then I moved over and began a role as a BEA SOA community development evangelist, and also ran the product marketing group -- the whole AquaLogic, WebLogic, and Tuxedo product marketing function.

Gardner: And you’re not affiliated with any particular vendor at this time?

Pendleton: No, I left BEA at the end of May to really look at SOA from a different vantage point.

Gardner: The first major topic of discussion is the convergence of SOA and open source. This is an interesting development. We have a number of different projects in incubation and being supported by a variety of different contributors, including some of the largest vendors in SOA and software infrastructure. These projects run the gamut from approaches to tools into runtime governance, enterprise service bus (ESB), and even the possibility of some frameworks. These include some complete frameworks, stacks for SOA within an open source environment, as well as promoting and developing a commercial open source business model, a variety of which we’ve seen emerge in the past several years.

We’re also going to discuss this week the topic of SOA and market evangelism, and that’s primarily why Jeff is here. The idea is that SOA is a concept, a methodology, a paradigm shift in the latest approach to IT, computing and applications. But the fact remains that there is no one entity that represents or sells SOA as a concept. SOA is being brought to market piecemeal, through standardization efforts, through commercial products, from a variety of perspectives -- not always with a lot of user input, or at least evident user input. And, so we want to discuss this whole notion of how SOA should emerge in the market -- whose responsibility is it to define it and promote it.

Let's get to the news around open source. I attended, just over a month ago, the EclipseWorld Conference in Cambridge, Mass., and talked with a number of folks here, including IONA Technologies, and IBM, as well as Eclipse Foundation, about a variety of different activities that are now under way, several of which are in Apache incubation.

So there’s not a whole lot of publicly available information about these activities, but let me just give you a quick rundown to demonstrate the breadth of what’s going on. IONA has an Artix commercial ESB, and they spun off an open source project from there called Celtix, maybe two years ago. They have now taken Celtix into the Apache Software Foundation, and joined it with Codehaus XFire, an open source Java SOAP framework, to create a new initiative, CeltiXFire (CXF).

CeltiXFire is aimed at providing an open source services framework, and it's going to join a number of other SOA-oriented open source projects, including the Eclipse SOA tools platform project. Also in Apache incubation is Yoko, a CORBA initiative. Also in Apache incubation is Tuscany, a services component architecture (SCA) initiative, and also Qpid, a message broker project that aims to define protocols -- in a sense take up where Java Message Service (JMS) leaves off.

So, there’s the possibility here of a rather robust and ecology-driven SOA infrastructure stack, that could play a role as an entire SOA approach for some enterprises, ISVs, and software as a service (SaaS) providers. Or bits and pieces of this could be cherry picked -- particularly, I imagine, from the tools perspective -- and used in conjunction with other commercial SOA products.

Let's take a quick level-set on this. How seriously should we take these open source SOA projects? Why don’t we begin with you, Neil?

Ward-Dutton: We should take them pretty seriously as people who look at the market. I think for too long the industry, the IT industry, and this community has not really looked hard enough at open source and freely offered technology products. It's time that we really should take them very seriously.

It’s interesting to think a bit about the distinction between open-source and free, and how much that matters to people, particularly to what you might call end-customers. Certainly, if you’re going to expand the definition to include free, but not necessarily completely open source, today you could just look at Sun Microsystems and what’s now called the Sun Java Composite Application Platform Suite (CAPS), which is basically SeeBeyond's old portfolio.

It's been reengineered to a degree to support SOA, and there’s a hell of a lot stuff in there, which is available at no cost whatsoever. As usual with the open source model, if you want paid support or enterprise support, you have to pay for that, but you can get the software for free. So, it's not just about Eclipse or Apache incubator software projects, it's also about the really big gorillas of the industry taking big steps toward making stuff freely available.

I don’t see why that should appear strange because this is all infrastructure. It's all boxes and wires, just plumbing, and increasingly that is commodity technology. We’ll get on in a bit to talk about the softer issues around SOA, and I think that’s where the real competitive differentiation, and the real challenge, actually come -- working out some of the more business-focused issues, and some of the more methodology-focused, and quality-focused issues.

Gardner: One of the things that I find interesting about the emergence of SOA, and the emergence of open-source SOA, is that this is happening very early in the development of SOA. We saw other open-source products, perhaps Linux, come a number of years after Unix and Windows were quite mature. And some people would even claim that there was some reengineering involved. Also the open source Apache Web Server followed the Netscape Web Server and Microsoft IIS Web server and became very prominent in the market, and so, again open source followed the commercial lead. What I think might be different here is that open source has an opportunity to lead the commercial vendors when it comes to SOA. Does that make any sense, Steve?

Garone: Yes, it does. I think what you’re saying makes a lot of sense, although I think it’s less of an issue in terms of the connection between open source and SOA, and more an issue of the timing of SOA. The reality is that open source, in the case of SOA, was around for a fairly long period of time. Whereas in the other examples you gave it wasn’t. So, you naturally expect the open source community to kind of jump on this new initiative, as it will with other new initiatives as they come along.

I agree with everything Neil said. I think it’s important to take this very seriously, which does not unfortunately imply that everything at this particular point in time has stabilized. One of the things that one likes to see when you talk about technologies and issues that are “open” is that standards bodies are sort of leading the way, and we see a lot of that. But, what ultimately happens, or what happens along the way, is that you get competition even on that level.

You see organizations like ObjectWeb and Apache, and others, going off and doing their own thing. In similar areas, some are supersets of others. But that introduces a level of confusion to the end-user. We hope that eventually that will shake out, but I don’t think in the case of SOA it has done so as yet. One interesting example of that, by the way, is the example you brought up of IONA and Celtix. Now, Celtix originally was hosted by ObjectWeb. I think one of the factors in that choice was the fact that IONA is a European company, and ObjectWeb is already active in Europe.

But now, what they’ve essentially done is move that technology, in the context that we talked about earlier, over to Apache Foundation. So, even in the case of one particular vendor, there wasn’t a lot of clarity over the last year or two in terms of where that technology would reside, and who would be leading the charge from the standard’s body point of view.

Gardner: Jeff, let’s go over to you. It seems to me another opportunity with open source is that the users, large enterprises -- something along the lines of General Electric, Ford Motor Co., General Motors, and Boeing -- can come in and say, "Listen, we want a seat at the table too. And we want to have a say in how these technologies evolve." So first, do you take these open source projects seriously, and is there something new and different here, where we are able to get end-user input, and an open source approach to some technologies that have not yet matured?

Pendleton: I spent the last couple of years at BEA, really looking at open source, first as a threat, and then one day realizing, as a lot of vendors have realized, that open source is actually a stimulus for change within these large enterprise software and hardware companies. When you’re sitting in the middle of these software companies you have these machines that you have to keep running, whether it's your maintenance streams, or your product streams, or what have you. And it’s very, very difficult to change those things without some sort of a very severe external catalyst.

So from my perspective, open source is innovating and causing innovation that the vendors are at first reluctantly but then wholeheartedly beginning to accept because they really don’t have much choice. But as I’ve talked to CIOs through the briefing center at BEA, what I‘ve noticed is there is a real fear -- and almost a deer-in-the-headlights phenomenon that’s occurring -- because there’s so much open source and there’s so much change.

What these folks are worried about is, first, they’re still a little bit in the woodshed from the last big, exciting IT wave. Second, they’re worried about the complexity of what they’re currently managing, and the complexity of all of these new ideas and standards that are coming out. To a certain extent open source is really having probably as much, if not more, impact on the vendor community to catalyze change there than it is in terms of really having a dramatic impact across the board in a lot of these enterprise companies. I think they’re selectively picking up open source components. But I think that, if anything, it's adding to just the churn out there. It’s an interesting dynamic, but as far as SOA goes, I agree open source really has been, arguably, a SOA poster child for years.

Gardner: In what sense?

Pendleton: Well, by the fact that we‘re talking about a lot of independent components that are really attacking very specific aspects of the plumbing of IT, doing it in a collaborative way, bringing the best minds together, and really providing a loosely coupled set of capabilities that then theoretically can be woven by architects and developers into a richer platform. I think that to a certain extent open source has been living the SOA loosely coupled federated approach for years.

Gardner: I suppose the only way that you could manage open source approaches is to divide things into small manageable chunks, and then develop those chunks in a way that they can be, at least to a large extent, dropped, like a LEGO piece, into an existing or heterogeneous environment.

Pendleton: What’s really cool about it is that you are able to draw people with passion and with a certain level of expertise into that specific LEGO, and so you get a much richer service offering. Then you look at the vendors, and they take that and blend it into their capability over time. So you really start to get those composite effects when you blend what’s happening in the standards world with what’s happening in the vendor world.

Gardner: Do you think we can take this another step further, which is to say that we might not even be at this level with SOA as we define it if open source were not such a major trend at this time? Would we still be looking at monolithic SOA stacks from specific vendors?

Pendleton: Are you asking if the movement toward a comprehensive stack for SOA would not be as far along without open source?

Gardner: Well, SOA has loosely coupled parts that use messaging and a variety of technologies across boundaries for purposes of integration and process refinement, and publish-and-subscribe capabilities. Would we even be at this point, in this openness, if open source as a trend were not an accelerant or a catalyst to promote that?

Pendleton: Having lived through a multi-billion dollar experiment at HP, I would say the reason e-services -- as a vision for the industry -- didn’t succeed back in the mid- to late-1990s was because we only had within our portfolio the existing monolithic, tightly integrated capabilities. There were no standards. I was trying to build basically an SOA, a billion-dollar SOA prototype, with no standards, no open source, really a very narrow gene pool to work with. What we ended building was another monolithic stack from a very poor source of ingredients.

So, yes, I absolutely agree that open source has in many cases created a richer gene pool for us to create the next wave of IT.

Garone: I would agree with that to a limit. I don’t think you needed open source to be able to move in the direction of openness and interoperability to the level that’s required for SOA. What I think open source might have done is sort of held back the tendency of vendors to add their own little hooks, and their own little proprietary extensions and so on that often take open standards and industry standards and make them a little more difficult to work with in terms of putting together a solution with components from a lot of different areas.

Gardner: Perhaps this is a three- or a four-leg stool, and SOA is the top of the stool. The legs that are supporting it over time, and we need all of them there, would be open source, open standards, and the move toward distributed Java computing over the last 10 years. Perhaps a fourth standard or pillar on this would be this notion of the extended enterprise. If we didn’t have the need for all of these things, or they weren’t in place, then perhaps SOA would not be as we know it. Does anyone want to react to that?

Ward-Dutton: I’ve been kind of scratching my head for a couple of minutes, so I’m glad you mentioned the last leg there, Dana, because I was going to jump in and say that there has to be a need for it as well. Maybe when you said "extended enterprise" computing, that captures that. Certainly SOA would not be here if people didn’t actually need it, and that’s actually the biggest hurdle that we collectively have to help people get over. It’s not primarily a technology question anymore. It’s actually convincing people that SOA is a good thing to do.

Gardner: It’s about the business.

Ward-Dutton: That’s exactly right. It’s about explaining why a SOA-approach to delivering IT capabilities is really a sensible thing do, given challenges: connecting organizations together, dealing with transparency, and compliance initiatives, becoming more effective -- that kind of cross-departmental business processing.

Gardner: That makes for a very nice natural segue into the next major discussion point today, which is this notion of evangelizing and educating on SOA. But before we go there, I want to ask one more question on open source in SOA.

I’ve heard several people theorize that the number and complexity, as Jeff pointed out, of the open source projects and componentry are going to be a bit overwhelming for your average enterprise. Therefore they predict a consolidation, and we’ve already seen that with the Codehaus and the IONA Celtix combination.

So what’s going to be the right balance between this componentization within an open-source environment of SOA and the need for simplicity, and a bit more integration and unification around a common approach, or at least a common methodology that might give you some choices among the components you choose or use? What do you think the market will require for this balance? And I will go to you first, Steve.

Garone: Well, that’s a really good question. And it’s a good question because it’s an interesting one for decision-makers. It’s also a very difficult one to answer.

For me, the criteria are not going to be so much: To what degree do we componentize versus have a more integrated approach for some application domains. To me the real issue is how do you transition an organization? Neil used the words "make the business case," and it’s a good thing to do.

It's also important to make the case that SOA is not a bad thing to do. And where that manifests itself is in: "What do I have installed today, and how do I best leverage that within the context of a SOA-based environment." To me, the answer is not so much how many components I use, how granular are they, and how much variety I have, but rather, "Can I make new application logic that I build within a SOA context componentize, and so on? Do I make that work well with what I already use, and what I already have installed, from the standpoint of applications, especially enterprise applications?"

Gardner: How about you, Neil? What’s your position on the proper balance between open source componentization in SOA, and an open source integrated stack approach?

Ward-Dutton: I think the best way to answer that question is to think about supply chains, about the delivery of software as a supply chain. The degree of componentization that makes sense depends on where you sit in that supply chain.

The degree of componentization we have historically had and what we are experiencing now is the reflection of two things. First off: The main people who are actually doing something with this stuff aren’t end-users, but they are really getting in there and making things happen. They’re still actually within the supply chain. They’re not at the end -- that’s one.

The other is that where it is an end-user situation that’s occurring where those people are playing, those are very leading-edge adopters of technology. And so there’s a high degree of tolerance for a high degree of componentization of these things. As the market matures I would fully expect to see much lower tolerance for that, and a much greater interest in aggregated offerings.

So you look at what SpikeSource does, for example, or you look at the announcement from Novell a couple of days ago around integrated stacks of Novell technology, IBM technology, open source, closed source. As you get into segments and scenarios, which are less leading edge, they actually want the reassurance that comes with bundling, and that comes with aggregation and comes with, "Don’t tell me about the details, just tell me it works."

It’s a question of maturity, of where you sit in the software supply chain. To a degree we're kind of dancing on the head of a pin, because -- don’t forget -- this is open source. Although there is a consumer predisposition to a certain level of granularity, actually it doesn’t really matter. Because within a project if someone wants -- metaphorically speaking -- to break something up within an aggregation they can absolutely do that because it's open. So, they can break something apart and mess about with it that way.

Gardner: Right. I would like to answer my own question through a little slightly different lens or focus, and that is around the politics, or the power politics, within infrastructure decisions. If I were an enterprise I might look to bring in open source components for those which I view as most likely to be a lock-in type of a technology. I might be more comfortable going with a commercial vendor on parts that I think will remain in some competition, and that I might be able to pick and choose among different commercial vendors as well as open-source components.

But for something perhaps like an ESB or a data-services approach or framework, and certainly from a tools perspective, I might be more inclined to go open source so that I can avoid lock-in and provide myself with a reduction of risk moving into the future, a future which I have very little insight into. Can you react at all to that, Jeff?

Pendleton: Well, that comes from a position where the architect is a key decision maker within the IT community. In organizations where architecture matters that makes a lot of sense. But my experience is that there’s kind of this interesting politic within IT communities where a lot of the open source decisions are still being made by what I call the "fighter jocks" that are in the development organizations.

To a certain extent they are bringing open source in because it makes their job a little easier, but also because it's cool. It's interesting to look at the political dynamic that’s happening right now within IT to figure out how to break that architecture-vs.-developer struggle that’s going on. So when you talk about bringing in choke-holds and things like that, it makes a lot of sense when the architects are in charge. But it’s a different kind of flavor when the development organization, which is being sponsored by a powerful line-of-business team, is really trying to drive an innovative piece of applications. I don’t argue your point so much as I think its only 50 percent of the equation. I think that a lot of the open source is coming in because it’s cool.

Gardner: Does a SOA by definition require a bit more coordination, communication, and harmony between the development teams and the architects -- and even the operators?

Pendleton: Architecture, and I mean the "A" in SOA, is really the defining element -- and the disruptive element. In my opinion that is going to really define the enterprise of the 21st century. And I’m not just talking about IT architecture. So, absolutely! I’m just pointing out that I think we don’t really have strong architecture in most corporations today.

Gardner: So people are often still thinking tactically about different parts of this but not holistically from a strategic overview?

Pendleton: When you talk about not wanting to get locked in it's generally something that’s more applied to vendor lock-in, traditional vendor lock-in, and it's generally more of a decision that’s being made by the CIO for very pragmatic reasons. But when it comes to open source there’s still a fair amount of the old West cowboy activities going on in most IT organizations.

Garone: If I could just jump in and add something briefly. That’s a great point Jeff, and in fact in my experience, especially over the last year or two, SOA within organizations has typically been done on a small scale, a pilot project and so on, which would be initiated in many cases by the development community by establishing a set of technologies. Maybe it's an ESB, maybe it's an application server, or open source in the situation we’re talking about here.

Then, when it comes time to expand this, when you have success, you want to expand it throughout the organization, and only then do the architects get involved. Now, you’ve got a technology and a basis of experience in a particular technology and there’s an incentive to some extent to stick with that. I think that’s really where what you’re saying comes into play.

Gardner: Well that makes another natural segue into our next discussion point. And that’s about how we evangelize and educate the concept of SOA as strategic, as architecture, as a long-term approach, when we’re still stuck in kind of a tactical decision-making process, and we have a variety of different vendors, and open source projects approach to componentry on the tactical level.

Let's start with you, Jeff. Perhaps you can explain some of your thoughts about evangelizing SOA and why you’re heading in that direction, both as defining the problem and then perhaps we can all think about some solutions.

Pendleton: Okay, as I mentioned early on I’ve been suffering from schizophrenia for a number of years, in the sense that I would be on the IT side really trying to drive application development. Then I’d jump over to the general management side. It was always amazing how my perspective on IT would change, depending on which hat I was wearing.

In one case, as IT, I was trying to bang the table saying, "We can really add competitive advantage here," and then moving over to the other side and looking at IT as an inhibitor and a road-block to achieving my quarterly numbers, and IT as something that was very, very expensive, and very risky.

So when I looked at SOA and talked to a lot of customers we were all in agreement that SOA was very, very disruptive, and in fact probably will be the defining element of the 21st-century enterprise. But for some reason we weren’t able to bring that message to the line-of-business guy, who still had this parochial, or somewhat skeptical, view of IT. They just weren’t hearing and understanding the potential.

So this summer we brought a bunch of folks together from some of the fiercest competitors, from some of the largest companies, to start talking about what is it about SOA that really is going to the define the 21st-century organization. I’m not sure we have the answer yet, but one of the things that we’ve noticed is that the vendor community is preaching to the choir for the most part. They’re talking about the power of SOA to folks who kind of understand the potential of SOA.

But when you go to the people that really need to see IT as strategic again we’re delivering messages like "agility," "flexibility," "adaptive," "liquid," which are like motherhood and apple pie. The real challenge is telling what it is about SOA that's so disruptive because I’m not sure line-of-business guys are going to care about SOA any more than they cared about client-server during the second wave.

What is it that is so disruptive? It’s not the LEGO approach per se because when I was at HP -- talk about a LEGO company; every LEGO was completely autonomous -- it was a cat-herding exercise. Loosely coupled was the HP way, and that didn’t work. Then you go to these large monoliths, where they have very much a top-down driven execution, and that doesn’t work.

So, what is it, and how do we start to make the potential of SOA a practical and pragmatic strategy? ... Make it something that IT people can bring to the business, so that the business people can say, "I get it, and what’s more, I can measure it, and I can see how it’s going to affect the various financial statements."

That’s what the industry is struggling with -- to break out of this preaching-to-the-choir marketing and really being able to take these very generalized notions of agility and LEGO-like use and translate them into something that’s consistent with what we saw with the second wave of IT, where we were talking about quality and the customer, and the notion of process core-competencies. We need things that business people can get their heads wrapped around, that are aspirational enough but also practical enough that they can execute over a couple of years.

Gardner: Let’s go to Neil. What sort of terminology do you think is required in order to bring to those line-of-business folks a business case for this fundamentally different IT approach?

Ward-Dutton: The challenge is that it's very difficult to say one thing that’s going to suit everyone, because SOA is kind of both everything and nothing. It's like dancing on a head of a pin.

Garone: Well, that message isn't going to work.

Ward-Dutton: Yeah, exactly.

Gardner: Well, that is the message right now.

Ward-Dutton: SOA, by itself, isn’t really anything. It’s just a way of thinking. It doesn’t do anything. You apply that way of thinking to a problem.

Just take a step back. The fundamental challenge we have today is that looking through five decades of “IT innovation” we’ve ended up in a situation where no one really knows where to go because the level of complexity in systems is so high, the level of brittleness is so high. Businesses change and changing requirements is very rapid.

The number of stake-holders involved in anything to do with IT is ridiculous. It’s a kind of locking up of the very gears that should help business and IT work together. And the reason we’ve got here is that business and IT together have kind of sleepwalked toward this situation because they didn’t have the ability to generate consensus between themselves that a short-term view had to be balanced with a long-term view.

You have to do both. What we have with SOA is a catalyst to start doing that. There are reasons and there are methods for actually keeping the short-term at the front of our minds, but also keeping the long-term in the picture at the same time. That comes back to be idea about architecture, but fundamentally it has to tied into the short-term because you can’t make something real that is about a "big" vision. People don’t buy it.

Even with business process re-engineering in the 1990s, it didn’t really work that well for everyone because they couldn’t sustain that level of commitment. There are a few poster children but it didn’t really happen broadly. So there have to be short-term tactical hooks, but at same time a much longer-term broader-value message just sits behinds it. And that’s the challenge we all face with SOA, because when you’re looking at the tactical side the message that an IT shop has to give to its “business customers” is absolutely tailored to that particular business situation.

That’s what I mean when I say SOA is everything and nothing. You can’t have one set of messages that works for both the long-term and the short-term story, and that works for everyone. Does that make sense?

Gardner: Yes, it certainly emphasizes the challenge.

Ward-Dutton: It didn’t give us your answer, of course.

Gardner: Well, you probably wouldn’t give it to us on this show, if you had it -- right? That’s probably a pretty valuable piece of information.

Steve, you and I have talked in the past about how risk reduction is a pretty good message around SOA -- that you can reduce your risk going backward in time, that you can leverage your investments and you can modernize and you can continue to deliver applications and services and data as people have been accustomed to.

But you can also reduce your risk moving forward, either into the foreseeable future where you know that you’re going to have certain requirements and a certain approach of agility, and you can reduce time-to-completion, and higher quality, and take advantage of a wider variety of services, and perhaps embark on some general reuse and object oriented principles. But you might reduce your risk even going beyond that into things you can’t predict and understand, or know yet.

But risk reduction is still also rather nebulous, and it's not something you can take to line-of-business people and say, "We’re going to reduce your risk. Can you double our IT budget?" Do you have any other thoughts here about the proper message that can be both tactical, strategic, IT, and business?

Garone: Wow! First of all, that was a great synopsis of the benefits and advantages of SOA. I think it’s right on. We’ve been talking here about the long-term versus the short-term, the tactical versus strategic and everything you just said, Dana, sort of plays into that.

To me the messaging, really the gap in the messaging, manifests itself in what you’re actually saying versus what you can actually demonstrate. It's the story versus the evidence. There’s no magic here. Vendors have for a very long time gone out to their prospects and customers with potentially very impactful and very interesting studies and case-studies that demonstrate ideally within their particular type of business how this technology has been used to create real advantages in terms of not only costs, but also in terms of business value and increasing the ability to do business more efficiently and so on.

To me that’s really the key. The messages are there, but the question is how do you demonstrate to a particular customer that this will work for them? For me, the only real way to do that is for vendors to work very, very closely with customers to make it happen for them in individual cases, and then bring that story to the rest of their customers.

Gardner: I think that sounds a little bit like a rehash of what we’ve been through for the last five years. After the dot-com bust, after the telecom bust, after the contraction in IT spending, everything was about ROI and TCO. And it sounds to me like you’re still trying to do an ROI-TCO cost-benefit analysis approach to SOA.

Garone: Certainly that's part of it but I wasn’t implying that it's all you need to do. Again, there are distinct advantages in terms of all the "... ilities" that you just went through that we recognize unto themselves, but may not be very effective to a business manager. If you then demonstrate those "... ilities" with real examples -- how it was done within their industry -- I think that can have a very positive effect. And, it’s not just restricted to numbers.

In fact I would make the case that not only should one not restrict yourself to that, but it maybe impossible to do anyway. A lot of the end users I know don’t really want that kind of information presented to people who are their potential competitors. Whereas in a qualitative sense they like the PR associated with being able to say we’ve done a really good job at this.

Gardner: Let’s get back to you, Jeff. Obviously we still need to work on the message. What about the messenger? We don’t seem to have a figurehead for SOA, like we’ve had in some cases in the past. We certainly had Linus Torvalds for Linux. We’ve had Larry Ellison for databases. We’ve had Bill Gates for client/server. Do we need a better messenger, in addition to a better message for SOA?

Pendleton: Well, I think there’s two sides to the coin. One is the messenger to the IT community, and then the messenger -- or messengers -- to the business community. The client/server, despite Bill Gates's evangelical power, was not what really catalyzed that wave. I don’t think line-of-business people bought client/server. They bought some higher level of abstraction. They were buying into the total quality movement. They were into this whole notion of core competency -- really understand your processes, and which ones add value, and which ones don’t ...

Ward-Dutton: ... And freedom.

Pendleton: Yeah. Go talk to SAP about taking over the processes that really don’t add value. That catalyzed the whole ERP wave, which catalyzed the client/server. There was this very interesting dynamic that met in the middle. What we don’t have right now are those inspirational, aspirational, definitional messages that define the enterprise of the 21st century.

Gardner: So we need a sort of a passion, and also a spiritual level to the drive to SOA?

Pendleton: Yeah, the Tom Peters way -- and that may be out of vogue. People may be very skeptical. But having watched HP spend billions of dollars executing strategies and then looking at the ROI after-the-fact, or looking at the ROI to justify a decision that had already been made, I think that we don’t yet have those agility strategies that will then drive the conversation with IT: When the executives ask, "So how are we’re going to do this?" and then IT says, "Well, it’s all about architecture, and look at what we’re doing with SOA." Somehow we need to marry what’s going on in the IT community with something equivalent in the business community.

Gardner: So, a complete narrative that brings in the tactical and strategic, and crosses the chasm between IT and business?

Pendleton: Yes, look at business intelligence. You can argue that it was catalyzed by Sarbanes-Oxley and all of the regulation, but the reality is that most corner-office executives really find out what’s happening in their business through interrogating their managers, and their managers' Excel spreadsheets.

How many guys are going to bed at night really wondering if they know what’s going on in their organization? The whole compliance-regulation thing has made it more powerful, but even without that as these corporations have gotten larger and more global it’s very uncomfortable to understand you don’t really have control. You’ve got this powerful enterprise with a lot of capability, and they want to know: "How do I bring that together, and bring a better offering to my customers, a richer offering? How do I get a bigger share of the wallet, taking advantage of my total enterprise?"

Gardner: Well you know who I need to invite on this show -- and I’ll go ahead and do it, but I’m not going to hold my breath -- Steve Jobs.

Pendleton: Absolutely. Everybody keeps coming back to that guy.

Gardner: If we could apply his genius of creating passion and drive around IT for its own sake -- never mind for these larger issues of strategy and ROI and TCO, which he usually doesn’t go to -- if we could get Steve to put his special sauce on SOA that might help a lot.

Pendleton: I bet he’d never even talk about SOA. I bet he’d never ever mention it, because that’s not what business people are buying. That just happens to be there to enable it but they’re buying something else, and that’s where we’re having the problem.

Ward-Dutton: Dana, it’s interesting the examples you gave when you were talking about evangelists. They were all really technical examples. Remember the point that -- I think it was Jeff -- made about people not buying client/server. I talked to people. Those guys were buying freedom. They were buying freedom from the shackles of the mainframe, which was something that was very closed, very powerful, and [freedom from that] held a lot of promise. A lot of people were very excited by it.

Gardner: It was mysterious. You never knew what was going on behind those glass doors.

Ward-Dutton: Right, and there was no way in hell you were ever going to get close to it. It was reserved because it was very, very expensive. In the mainframe era people bought dependability, efficiency, and accuracy because it was all about very high volume and very reliable transaction processing.

So that’s what people bought. They bought a tool to help them automate things that had to be done quickly and by rote, very, very dependably. In client/server, they bought freedom. They bought the freedom to use what was a radically different model of IT to devolve power and responsibility to individual lines of business. So it didn’t have to depend on those weirdoes sitting in the data center.

With e-commerce they bought into a promise of global markets, and the ability to radically change business models. SOA gives us all of those things, but the problem is that the IT community has created its pitch, as they would say over here. They’ve cried wolf too many times over their promises. Why should they believe us now?

The problem is that SOA actually does all those things. It helps organizations by aligning IT and business more closely. It helps organizations become more service-oriented and componentized, which enables organizations to more quickly adapt to new business conditions, enter new markets, compete globally, and do all that kind of stuff. It helps them be rock solid and dependable in how they get hold of information, so they know where they are. If you don’t know where you are, you can’t innovate.

Gardner: Let’s look at this lineage of freedom. We had the freedom of the PC, which then bled over into the freedom of client/server, which then bled over into the freedom of the web, which then bled over into web services, and is now bleeding over into SOA.

So, it's all about individual empowerment and freedom, which comes at a very auspicious time because this "echo boom," that is to say, the children of the baby-boom generation, are now thinking quite a bit differently. They like independence. They do not like a top-down approaches. They do not like to be told how to do things. They like to explore on their own, as evidenced by such things as Web 2.0 and social networking, and MySpace.

To appeal to this younger, echo-boom generation, we need to encourage that level of exploration on your own terms, individual approaches, not a cookie-cutter kind of methodology, but a kind of a messy, creative-chaos approach. So, do we take this continuation of freedom, and apply it in some better way to what services on your terms is all about? Jeff?

Pendleton: Well, freedom is great but when you have to hit your numbers every quarter, when you’re signing up for a two-year plan, you need to have confidence that you’re ultimately going to execute on a portfolio of strategies that execute against those management-by-objectives (MBOs). The MBO is what drives line-of-business guys.

I think that you'd want freedom to invest and enable some percentage of your strategy, but at the end of the day we’re kind of handcuffed by this need to hit our numbers every quarter and every year. I’m not sure how that’s actually going to manifest over time, but the last thing I would want to do is have more cowboys in my IT organization.

What I’d like to try to figure out is how to have a blueprint, and a strong architecture team ... imagine if you were building a corporate headquarters. You probably wouldn’t want to do that with just a free-wheeling set of cement layers and plumbers, and what have you. You’d probably want to have a really rich architecture and blueprint.

Ward-Dutton: It's about who has the freedom, isn’t it?

Gardner: But wait. I read about this guy, Frank Gehry, who builds buildings that look like they were put together by creative chaos, but are actually highly architected in order to make that possible.

Pendleton: But are there a thousand developers involved in actually building that blueprint. He draws the picture, and then he works with sub contractors and contractors to execute on that.

Gardner: Do you think this is a valid metaphor for SOA, a Frank Gehry-type architecture?

Pendleton: Absolutely! Positively! I was looking at a picture of a rabbit warren. I don’t know if you’ve ever seen one of those, but you take a bunch of rabbits and you put them on a 100-acre field, and basically you build the enterprise of the 20th century. It's just holes going everywhere. They just dig, dig, dig. It's kind of is a picture of what the architecture of most enterprises look like. We don’t have the same type of pride and passion going into the internal architecture and workings of the enterprises that we do going into the corporate headquarters.

Gardner: That very interesting, isn’t it? Perhaps from the IT architecture perspective we've got a Frank Gehry-looking application that actually is as messy inside as it is outside. But what we really want is to have an elegant creation that’s beautiful on the outside and the inside.

Ward-Dutton: It’s about the right freedom for the right people. You can’t build a creation in enterprise IT because change is too constant. So you can’t build something. You can’t build anything. All you can do is build scaffolding.

Gardner: You can build an architectural approach, right?

Ward-Dutton: Right. A great approach to architecture is to partition a business and to partition IT up into domains which have different qualities. Some of them will be very stable, and very predictable. And for those you can actually optimize for those things you may not use a SOA approach.

In certain areas, particularly around the edge of the organization, where you want to connect with customers, with partners, with suppliers -- where you really want to innovate in terms of how you build products and services, that's for SOA. Actually, you don’t want to build a building. You don’t even want to build an open-plant building, where you can move partitions and cubicles around, although that’s better.

What you actually want to build is foundations and scaffoldings, so that prefabricated chunks can be pulled together very, very quickly, and torn down very quickly. That’s not about building a building, it’s about building the tools to build and tear down bits of buildings very, very quickly. And in some places you want to do that and in some places you don’t. It’s about balancing the two.

Pendleton: Sounds like a Hollywood sound stage.

Ward-Dutton: For some bits of a business, absolutely. Not every thing.

Gardner: All right, gentlemen, I’m afraid we’re going to have to leave it there. We’re out of time, and obviously this is a very interesting subject that we’ll be revisiting over and over again. And perhaps we’ll get closer to how to best conceptualize, and go to market and provide messaging for SOA.

So I’d like to thank our guests, Steve Garone and Neil Ward-Dutton, as well as Jeff Pendleton. I’ve been your host Dana Gardner, principal analyst at Interarbor Solutions. You’ve been listening to a weekly Briefings Direct SOA Insights Edition. Come back and listen again next week. Thanks everyone.

I would also like to point out to our listeners that if you'd like to learn more about BriefingsDirect B2B informational podcasts or to become a sponsor of this or other B2B podcasts, feel free to contact me, Dana Gardner, at 603-528-2435.

Listen to the podcast here.

Transcript of Dana Gardner’s BriefingsDirect SOA Insights Edition, Vol. 3. Copyright Interarbor Solutions, LLC, 2005-2006. All rights reserved.