Thursday, November 06, 2008

Semantic SOA

So was talking to my friend Benjamin Grosof, ex MIT Sloan Professor, from Paul Allen's cutting edge research company Vulcan Inc. While talking about next generation stuff he introduced me to the concept of Semantic SOA, he personally believes this will be the next big thing. It took a while for the concept to sink in, but once it did I was shocked to my core over the possible impact of such technology. Unfortunately he's twisted my arm, so I can't talk about it further :( But I'd be very interested to here what other people think this is about, and whether Gartner or other analysts have picked up on this yet.

update (7th of Nov 2009): heh, this was a hoax for a bit of fun. While at BRF I mentioned that I'm pushing the term "Business Logic" for a product that was neither rule centric or process centric. So our minds went wandering about other possible names, then obviously we thought it would be funny if we combine the hype of Semantic Web with the hype of SOA and see the response :)

5 comments:

  1. I love the idea. I spend my life designing and implementing SOA solutions for EAI and automated business processes. We have a long-running corporate joke in our company about our plans to design and create the 'psychic adapter' - the EAI adapter that will allow our SOA message broker to connect to anything, anywhere, using whatever protocols are required, dynamically selecting the correct addresses and bindings, implementing whatever message exchange patterns are necessary, applying whatever reliable messaging and total ordering approach we need, negotiating suitable retry policies, fallbacks, throttling, etc., automatically configuring service windows and, of course, performing whatever decoding and encoding, message-level security, validation and transformation processing that is needed. It will automatically understand trust boundaries and apply single single-on and other forms of identity management and federation as required, just as it will automatically understand the concept of parties and roles in cross-organisational interchange. It will, of course, understand and apply the tracking and KPI requirements for each endpoint and will dynamically route messages on to where they need to go. The point of the psychic adapter is that, once deployed, it won't require further configuration of any kind, or any additional development or administrative effort...it will just know!

    I just mentioned the idea of Semantic SOA to a couple of my colleagues, and their immediate response was - 'hey, it's the psychic adapter'! Personally, though I would love to be proved wrong, I won't be holding my breath on this one. SciFi SOA, perhaps?

    ReplyDelete
  2. I have a quite biased opinion since I'm working on this stuff but I still think it's worth giving a few pointers.

    The Semantic Web Services community has been pursuing this idea for a while, although it is more recently that people start to talk about it in more general terms as "Semantic SOA".

    You might be interested about a few European research projects such as DIP, SEKT, SUPER or more recently SOA4All that are based on these ideas. Derived from these projects it is worth mentioning the WSMO family of languages and the so-called Semantic Execution Environments that implement the vision, namely WSMX and IRS-III. (These are indeed not the only work in the area, just those that I know better.)

    Part of this work is under standardization within OASIS by the Semantic Execution Environments Technical Committee. In fact, we are about to release a first document for public comment which would benefit much from your input.

    From the technical side, myself I have recently started integrating Drools within a monitoring tool that is deployed in what one could safely call a Semantic SOA. The new features you guys have added for CEP are quite appealing in this sense. The lack for an out-of-the-box support for ontologies and their languages is a bit of a drawback but using JavaBeans one can avoid the issues. Despite the relative early stage of our work in this respect, I believe Drools will play an important role in implementing the Semantic SOA vision within our current work, especially now that it has support for CEP. Keep up the good work!

    ReplyDelete
  3. I find it ironic that some "introduce" folks to the concept of Semantic SOA (SSOA)! I've been designing and evangelizing the concept of SSOA since October 2004.

    I've presented the concept at numerous public forums such as the Semantic Technologies conference, International Standards Organization Open Forums, Object Management Group, Sun Microsystems Jini™ Community Meetings, DODIIS Worldwide conference, JavaOne, etc. I've co-authored published works in Ontology Lifecycle Management and Semantic Service Oriented Architecture.

    My personal website at www.semsoa.com basically includes my (joint) thesis and a white paper overview of SSOA.

    I lead three R&D programs under the banner of SSOA (two while I was still in the Navy and one as a defense contractor at SAIC.)

    Why there's even a Wikipedia page for SSOA! And if you look closely, you'll note the content of it aligns with my white paper. Of course, the Wikipedia entry needs further development, so please feel free to contribute!

    (Alas, all the good stuff related to SSOA is not publicly available though! As much as I like open source, I'm still a business man. and if you know any GREAT developers looking to work in this field, please send 'em my way!) :-)

    Lest I sound boastful, I'm not the only one; I've read others' published works that explicitly name SSOA.

    My point? 'It aint new!'

    P.S. I argue "semantics" aren't a modifier, enhancer, or variant of SOA; semantics are foundational and fundamental. without them, it will be exceedingly difficult, if not impossible, to truly overcome point-to-point IT architectures.

    P.P.S. Please pass my regards to Benjamin. We met at the International Semantic Web conference in Hiroshima, Japan.

    ReplyDelete
  4. update (7th of Nov 2009): heh, this was a hoax for a bit of fun. While at BRF I mentioned that I'm pushing the term "Business Logic" for a product that was neither rule centric or process centric. So our minds went wandering about other possible names, then obviously we thought it would be funny if we combine the hype of Semantic Web with the hype of SOA and see the response :)

    ReplyDelete
  5. LOL. But then the joke has rebounded on you Mark. The term 'Semantic SOA' is already taken and has a life all of its own! It is all about combining the Semantic Web with SOA. There are open-source projects like WSMX, which are delivering semantic web service execution environments, together with a number of commercial and military projects to establish semantic SOAs. There is even some work going on via OASIS to create standards for this stuff - e.g., a reference ontology.

    I took a very cynical stance earlier in this chain because of the obvious level of hype which surrounds the concept. I've just had reason to revisit UDDI, and was struck by the way in which UDDI is, in large part, all about capturing service semantics. Then I remembered that exactly the same sort of hype surrounded UDDI when it first came along several years ago! There were very similar claims about what it would enable. UDDI has an important role to play in SOA, but never even began to live up to the hype. However, taking a less cynical stance, I can see that the combination of the semantic web with SOA, enabled through frameworks that allow us to capture and process service semantics, could, in time, prove to be really useful in mainstream SOA design and implementation.

    ReplyDelete