SPARK: Exploring the convergence of Web 2.0, SaaS, and SOA
posted Monday, 20 March 2006
We wrapped up the SPARK architecture workshop late yesterday afternoon here in sunny Las Vegas. This very intriguing event on the future of software architecture consisted of a two day conversation between some of the brightest people in software. The goal: To try to figure out where it's all going. The trends that are generally associated with the concepts of Web 2.0 and Software as a Service (SaaS) are really reshaping views and expectations, while the industry's experience with SOA up to this point has been mixed. Web 2.0 and SOA, and to a lesser extent SaaS, are first order collections of software architecture forces and contraints that overlap quite a bit. But there are also clearly disjoint concepts like Web 2.0's social aspects and SOA's orchestration and its galaxy of obscure standards. The convergence of major visions of software architecture is a fascinating and vitally important topic, as I discussed in my cover article on the subject late last year in the SOA Web Services Journal, Web 2.0: The Global SOA. Since then, the blurring of the Web and the enterprise is something I've started calling Enterprise Web 2.0.
In any case, the event was sponsored by the Architecture Strategy team at Microsoft, though I think it's actually pretty important to point out that Microsoft's products and company strategy were not topics of discussion over the weekend. To their great credit, the entire two days of conversation roamed purely across the intellectual terrain of the future of software architecture.
The Participants
An extremely impressive line-up of folks from across the spectrum attended that ranged from people directly in the middle of the Web 2.0 phenomenon (like MySpace CTO Aber Whitcomb who told me that they are currently adding 250,000 new users a day) to well-known thinkers in the SOA space like David Linthicum and Anne Thomas Manes. Other notables included the original author of DHTML, Scott Isaacs, Google's Gregor Hohpe, ZDNet's Phil Wainewright, and Walt Disney's Steven Davis. Microsoft had a relatively minor presence but included a number of notables, especially two key folks involved in SPARK, Michael Platt and John deVadoss. Even Microsoft's uber-blogger and Web 2.0 Workgroup member Robert Scoble dropped into the event for most of day two. Dare Obasanjo and I never did get to have our own cage match but we had a really good discussion at the SPARK dinner that evening at Aquaknox.
What Happened
Of course, two days isn't nearly enough to tackle the monumental topics of Web 2.0, SOA, and where software architecture is going but we really gave it the old college try. There was a lot of brilliant conversation through the first day as we tried to identify the key drivers that are shaping software these days. The whole list was very extensive and discussed Web as Platform, ubiquitious computing, the drive to mass serve micromarkets and much, much more (see pictures below). It was tough to come up with a complete list and I'm sure we missed a few but I think you'll agree that we got well into the ballpark. The results were then used to bound the discussion on day two.
On day two of SPARK we broke into groups and attempted to create models that captured the drivers we had come up with the previous day. Given the size and complexity of the problem space, I can truthfully say it was a challenging task but a plenty of interesting material did come out. Just one small example: Anne Thomas Manes, Gartner's Nick Gall, and others at their table came up with an interesting Celtic Tree of Life "mashup" vision for how software and business ecosystems function. At the end of the day, we gathered up all the models and created a new very high-level master model facilitated by Michael Platt. It was obviously not easy to determine which elements were central and which were secondary but I think it's a good first cut. We're going to present the results this afternoon at MIX 06.
I could write for hours about the exciting discussion and ideas that came out of SPARK and indeed I intend do some more in-depth posting over the next week as I digest what happened and zero in on some of the more important individuals discussions. Also, since the entire conference was in the public domain, I was able to podcast all of Day One and the morning of Day Two for posterity and I'll post this as an amendment to this post once I convert them. I also took lots of pictures during the entire event, which you can find on Flickr here. Finally, a lot of the other SPARKers have blogged about the event, here are the links.
Blog Coverage of SPARK
I spoke to Microsoft's John deVadoss after the event and he is very interested in continuing the conversation since there's obviously lots more work to do. Expect to see more events like SPARK as Microsoft's really tries to get a handle on where software architecture is going. I commend them for bringing together such a terrific group of folks and essentially turning them loose to be creative and forward thinking. Also, Incite Partner's John Burgess and Dan Caine did a world-class job as the event facilitators. They apparently helped out with the vision for MIX as well and I think we'll be seeing a lot more of their work in the future.
Update: Mike Platt, Anne Thomas Manes, and myself hosted a session at the end of the first day at MIX that started communicating the results to the greater community. Microsoft is definitely trying to be open and inclusive with these efforts and they're making all the SPARK materials public. Mike did most of the presenting and Anne and I were only there for the discussion piece at the end. But it was interesting to note that Tim O'Reilly even stopped in to see the results of what 28 top IT architects thought of the direction of software architecture. Below are two slides that capture the important pieces: the problem space and the architecture model that resulted. The problem space and the architecture diagram needs a good amount of explanation and I'll try hard to get in a post soon that does just that.
What do you think? Will experts help drive the future of software architecture or is it just too community-powered now? links: del.icio.us
250k new users a day, that is scary! I remember a few months ago people
at myspace were adding a couple of servers a day, i wonder what they are
up to now.
