Hi Anne, Thank you this is very useful.
I was very careful to point out that "the author claims IBM equate SOA to EAI hub-and-spoke and ESBs are the hub." And I never claimed that IBM made such an assertion. In my blog I ask (rhetorically) IBM to refute this. I have some ex-colleagues/friends that work at IBM and they claim that there is so much to sell that IBM are confused over what to offer. When I talked about some vendors "putting lipstick on the pig" (i.e. EAI vendors slapping Ws on front of their hub ... or "message bus" vendors pushing their bus with some WS interface) on IBMer claimed "the problem is were beyond that. We've put lipstick on 20 and we say find he pig". What annoys me is that IBM has so many good products and so many good people but they still get all caught up in this FUD cycle. As you point out it wasn't long ago that they said ESBs are a pattern and not a product. I think that was FUD until they decided where they could apply lipstick. :-) It is my opinion, and we've spoken about this at JavaOne, that when the big players find that they are behind, say in 3rd or forth place, they create FUD to slow the market adoption down. If they find that they are not likely to catch up then they question the validity of the market and invent something "new". They change the market. Leaving the other vendors holding the "old" bag. That's not to say that the innovations are bad - the innovations bring some great improvements. IBM's policy of supporting products "forever" is very comforting for many of their customers. But it can also hinder IBM from having a coherent product story. I think that is what's happening here. They have too many products and are unwilling to focus. On the other hand they are making lots and lots of money. :-) They have a different bag of products for SOA for every customer. Each bag leveraging the existing IBM assets and architectural approach at each customer. What can you say to that? And with IBM GS applying SOA "know how" (and running interference) how can you compete on a Big Blue site? (rhetorical) Thank you for replying and giving us your insight. Your knowledge of their products is impressive. (As it should be), Regards, William On Dec 15, 2005, at 4:39 PM, Anne Thomas Manes wrote: > Okay, Henry, I will attempt to provide a voice of reason. Perhaps > Jason will chime in, too. (Ron's supposed to be on his honeymoon, > so I don't think he will.) > > John is spot on with this observation: > > I think that at some level, IBM sees SOA as a branding glue for > their disparate and overlapping software offerings > > IBM presented its "SOA Reference Architecture" to us, which was > basically an architecture that allowed IBM to position how about 40 > of its thousands of software products address the requirements of > SOA. Never forget: IBM's goal is to sell products. Therefore all of > IBM's marketing messages have a product focus. > > I think perhaps Phil Howard was more impressed than I was by IBM's > story at their annual analyst conference (held in Rye Brook, NY, 30 > Nov - 1 Dec). But I think we came away with a similar sense of > IBM's strategy. IBM has fully embraced SOA, and SOA is the focus of > all IBM software marketing messages. (See Phil's first article in > the series, "IBM Mainlines on SOA": http://www.regdeveloper.co.uk/ > 2005/12/05/ibm_soa/.) > > I do have a bone to pick with Phil, though. In his fourth and final > installment in the series (see "The problem with IBM's SOA > message": http://www.regdeveloper.co.uk/2005/12/14/ > ibm_soa_message/), he says: > > "The first and most obvious problem with IBM's take on SOA is that > you cannot simply go to IBM and say "I want to implement SOA in > this part of my business, what product do I need?" Or, rather, you > can say that, but the answer you get back will be horrendously > complicated." > > As Jason has so eloquently said, "SOA is something you do, not > something you buy". So the idea that you should be able to go to > IBM and get everything you need to do SOA is fundamentally flawed. > But perhaps I'm misinterpreting Phil's point. He does redeem > himself later in the article, when he talks about the need to bring > in IBM IGS to help with the cultural issues of SOA. So let me give > him the benefit of the doubt and assume that his issue with IBM's > strategy has something to do with the fact that IBM has 4000+ > software products, many of which overlap in terms of functionality. > > Take, for example, IBM's ESB strategy. IBM provides two ESBs: > WebSphere ESB (simple) and WebSphere Message Broker (advanced). By > the way, they are based on completely different code bases. > > Or IBM's mediation strategy: the 2 ESBs, WebSphere Process Server, > IBM Tivoli Composite Application Manager for SOA (ITCAM), the > DataPower appliances, Tivoli Federated Identity Manager, (and I'm > sure there are half a dozen more). > > It's nuts. And as Phil says, much too complicated. Now, you do need > multiple types of mediators in a SOA, but IBM really should have > coordinated its efforts and have its various mediators work as if > they come from the same company. (Shockingly, the products that are > best integrated with the others are the DataPower appliances -- the > only products developed by a third party.) > > But back to Henry's concern: Phil's assertion that SOA equates to > EAI's hub-and-spoke architecture, and that ESBs provide the hub. > (See "More on IBM and SOA": http://www.regdeveloper.co.uk/ > 2005/12/06/ibm_soa_comment/) > > Note that IBM never told us that SOA is a hub-and-spoke > architecture. This is just Phil's interpretation of IBM's product > architecture slideware. IBM, like most ESB vendors, graphically > depicts an architecture in which the ESB provides the core > communication infrastructure for the SOA, and it implies that all > messages flow through the ESB. Therefore it's easy to misinterpret > the image and come away with the sense that the ESB acts as the > central hub of the environment. > > But as I said before, IBM's architecture is all about product > positioning, not about SOA architecture. IBM is still a little > conflicted by what it means by ESB, too. Keep in mind that until > September, IBM was still referring to ESB as an "architectural > pattern", not as a product. Now that it has two products that it > refers to as ESBs, IBMers sometimes get a bit confused by their own > architecture. Quite a few IBMers still think of ESB as an > architectural pattern, and the ESB products are just one of many > products used to implement the pattern. But then you talk to the > product marketing people, and they, or course, want to map real > products to the architectural pictures. > > But regardless of IBM's state of confliction, I heartily disagree > with Phil's assertion that SOA is the modern-day equivalent of EAI/ > hub-and-spoke. On the other hand, most ESB products (not including > IONA Artix, mind you) *are* the modern-day equivalent of EAI/hub- > and-spoke. And I'm sure you'll be shocked (!) to hear that I am not > a huge fan of ESBs -- for just that reason. (See my blog on the ESB > market: http://atmanes.blogspot.com/2005/04/future-of-esb-market.html) > > Nothing about SOA imposes an architectural requirement to route > messages through a central hub. (Interesting aside -- IBM's new > Service Component Architecture [SCA] has no requirement for a > central hub, either.) I strongly recommend using mediation to > manage message traffic and to implement policy enforcement points, > but mediators are often scattered throughout the environment, and > they often execute at the endpoints. So the idea that SOA = hub-and- > spoke is very misguided. > > Anne > > On 12/14/05, John Kupyn <[EMAIL PROTECTED]> wrote: > IBM claims, to its partners at least, to have 300 patents related > to SOA. I think that at some level, IBM sees SOA as a branding > glue for their disparate and overlapping software offerings (a > point made before). However I also think that a everyone at IBM > would agree that "SOA takes advantage of autonomous distributed > services collaborating together in business processes." > Stay strong. > JK > > William Henry < [EMAIL PROTECTED]> wrote: Okay so I read > another article form the same author as my last post > about IBM's strategy/story. > > in it the author claims that IBM equate SOA to EAI hub-and-spoke and > ESBs are the hub. > > Article here: > http://www.regdeveloper.co.uk/2005/12/06/ibm_soa_comment/ > > My comments here: > http://www.ipbabble.com/2005/12/soa_does_not_equate_to_eai_hub.html > > I hope all is not lost. It's enough to drive a person to drink .. and > it's only Wednesday. > > Can we have a voice of reason here from Anne ( T.M.)? Any comments on > this Anne? > > Regards, > William Henry > > William G Henry > [EMAIL PROTECTED] > http://www.ipbabble.com > > > > > > > Yahoo! Shopping > Find Great Deals on Holiday Gifts at Yahoo! Shopping > YAHOO! GROUPS LINKS > > Visit your group "service-orientated-architecture " on the web. > > To unsubscribe from this group, send an email to: > [EMAIL PROTECTED] > > Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service. > > > > > > SPONSORED LINKS > Service-oriented architecture Computer monitoring software Computer > and internet software > Free computer monitoring software > > YAHOO! GROUPS LINKS > > Visit your group "service-orientated-architecture" on the web. > > To unsubscribe from this group, send an email to: > [EMAIL PROTECTED] > > Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service. > > ------------------------ Yahoo! Groups Sponsor --------------------~--> Most low income homes are not online. Make a difference this holiday season! http://us.click.yahoo.com/5UeCyC/BWHMAA/TtwFAA/NhFolB/TM --------------------------------------------------------------------~-> Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/service-orientated-architecture/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
