Manes: in today's climate, be specific about SOA-related requests

Manes: in today's climate, be specific about SOA-related requests

Summary: In the second half of our recent BriefingsDirect podcast, hosted by Dana Gardner, we heard more from Anne Thomas Manes, who provided more perspective on  her recent proclamation that SOA -- at least as we knew it -- is dead. (Link to podcast here, full transcript here.

SHARE:

In the second half of our recent BriefingsDirect podcast, hosted by Dana Gardner, we heard more from Anne Thomas Manes, who provided more perspective on  her recent proclamation that SOA -- at least as we knew it -- is dead. (Link to podcast here, full transcript here.)

The key factor is that in a rough-and-tumble economy, if you want money for a project, you have to show the business the money. As Anne put it:

"If you go before a funding board this year -- if you are an IT group and you are trying to get funding for some projects -- and you go forward with a proposal that says we need to do SOA, because SOA is good, it’s going to get shot down."

Instead, she advises, go to business decision makers with "very specific value-add projects" that use service oriented principles to get the job done quickly and efficiently.

What will work in 2009:

"'We need to build a billing service which replaces the 27 different billing capabilities that we have in each of our product applications out there.'"

What won't work in 2009:

"We need to go get an ESB. We need to go get some registry and repository technologies. We need to invest in all the SOA infrastructure. We need to do SOA just because SOA is what everybody is telling me we need to do."

Anne's point -- as she was making in her original "SOA is dead" post -- is that too many companies embarked on SOA efforts for SOA's sake. SOA is the mechanism, a philosophy if you will, that can bring business solutions to reality. But SOA by itself won't make a dime for the business.

Tont Baer -- also on the panel -- also provides insights on our recent discussion with Anne, noting that she updated "the classic argument that the business wants solutions that address business problems and deliver business value... Since the popping of the dot com bubble, few enterprises have had any appetite for technology projects justified on the basis of technology."

Topics: Software, Browser, Enterprise Software, Software Development

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

9 comments
Log in or register to join the discussion
  • Can we apply basic logic?

    Joe,

    I am not sure why this debate continues. Is it fair to report on whatever somebody says even if it does not make any basic sense?

    >> ?We need to go get an ESB. We need to go
    >> get some registry and repository
    >> technologies. We need to invest in all the >> SOA infrastructure. We need to do SOA just
    >> because SOA is what everybody is telling me
    >> we need to do.?

    Anne recommends to focus on "building services". But how can you do that if you don't have a "service container"? how can you do that without what a service container (a.k.a ESB) has to offer? exception anyone? versioning? logging? virtualized endpoints? backend connectivity? transport independence? transformations? orchestration?

    Did Anne ever built a service in her life?

    Assuming you built a service without a service container, who's going to find it? Now, I don't recommend anyone spends more than 100k for their service registry (all inclusive). I have seen horror stories where products that simply don't work and cost a real bundle (500k-2M range...).


    I am actually surprised that in 2009 somebody can still be reported on and not pass basic logic tests. Considering how much damage Anne's comments have made (yes it has impacted my SOA initiative already, a smart ass mentioned that he had read in the press and at Gartner that SOA was dead and suggested to pause my project). If Anne's goal was to disrupt countless project worldwide, well she sure has done really well. Congratulations Anne ! Thank you Joe, for keeping this useful thread alive.

    As far I as I concerned she has lost any and every credibility she could have ever had on SOA.

    Just my 2c.
    jdubray
    • I agree on the harm of "SOA is dead"

      Manes really did a disservice with that article. There were many posts about how the title didn't match the content, how her intent was in the vein of "The king is dead; long live the king" but given how many people read just headlines and rarely the Cliff's notes (let alone the actual content) it seemed inevitable that legitimate efforts would come under fire for this fairly flippant article title. Shame on Manes.

      "Anne recommends to focus on "building services". But how can you do that if you don't have a "service container"? how can you do that without what a service container (a.k.a ESB) has to offer? exception anyone? versioning? logging? virtualized endpoints? backend connectivity? transport independence? transformations? orchestration?"

      You make it sound like all of these must be completely new or redone. These are all implementation details that can be addressed in a wide variety of ways. SO implementations do not require huge investments nor complete retooling. It's highly likely that most companies already have the pieces they need. They just need to use them in an SO fashion--after having defined the architecure of course.
      reamon@...
    • Self-fulfilling prophecy?

      Wow -- talk about the risk of self-fulfilling prophesy. You see the same issue with the economy -- the more doom-and-gloom talk there is, the more people get scared and pull back, making things even worse. You're right in that the "SOA is dead" tempest has run its course... and I'm sure someone will be posting an article a year from now, five years from now, declaring SOA dead again.

      But I think what Anne was trying to say is not that the principles of SOA should be abandoned, but that too much hype got wrapped around the buzzword itself. And the fact that "pure" SOA doesn't exist -- rather, companies need to pursue a range of strategies.
      joemckendrick
      • Anne knows exactly what she is doing...

        Joe:

        listen, the last thing you can say is that Anne is stupid enough to not see what she is doing. So maybe one day, people will only pay attention to analysts when they speak reasonably and clearly about their analyses.

        The same analysts that a few years back sent people doing these big top down SOA projects are now telling us, to "start with services". Maybe, we could just stop this insanity... that would be a big help for those of us that actually do work.
        jdubray
  • Good advice in any climate

    I have a lot of respect for Manes. Her analysis and explanations are always worth a read.

    I'm really glad that she's writing more about "don't sell SOA" but she's a bit late to the party. Selling SOA (or OO, or EAI, or any of the approaches and technologies we've seen over the years) has always been an ill-advised notion. A business-focused, business-value approach is always the right way. SO for SO's sake is bad mojo.
    reamon@...
  • RE: Manes: in today

    Yes, I agree that there are a lot of pieces already in place at most companies and a lot of these pieces do not cost any money.

    The argument I have with Anne is that recommending to focus on "services" without thinking how these services fit in an architecture is simply silly. So telling people, if you don't have the SOA building blocks, don't focus on them, just go an build services is possibly the worst recommendation she can give to these people. SOA requires that you think more (without spending a ton of money) rather than less.
    jdubray
    • Agreed

      Great point. SOA is first and foremost about architecture. Building services without a guiding architecture is problematic.

      I think the key here is defining an architecture with a view towards addressing business items in business terms and building out the services defined by the architecture that provide a business benefit.

      The effort must always have a business viewpoint. Those that control the purse strings don't care about architecture, SOA, REST, etc. so never, ever seek funding for those parts of the effort in isolation. One must have an architecture but don't put it as the main point in the funding request.
      reamon@...
  • RE: Manes: in today

    If you go into a budget meeting and say "We need to migrate to an SOA environment", you are selling the wrong thing.

    If you walk into a budget meeting and say "we need to upgrade the HR software to work with this payroll and allow for us to have a sitewide managed login" you have a chance.

    Note the second does not say anything about the underlying architecture. Leave that to the architects. Note that it also is avoiding a lot of the buzzwords which translates to dollars such as "transformative". You aren't out there selling an architecture, you are selling what is delivered. You need to specifically address the service that is being transitioned and how it will integrate with the current systems because not all of them will be moved.

    The underlying architecture just is not necessary to discuss at this point. Present a workable plan and sell that. The reality is that a lot of companies did implement SOA because it was the latest thing and right now a lot of consulting companies are selling SOA services that are SOA in name only.

    When it comes time to implement, then you can do SOA to your heart's content. But, the reality is that what we call SOA is just going to get rolled into best-practices and will lose a separate identity. It will just be expected that the architecture has certain features.
    rijrunner
  • RE: Manes: in today

    Guys, at the end of the day, no matter how you look at it, tons of very valuable services are going to emerge from the Cloud, not just from your organization.

    I really like the insight that Matthieu Hug had about SaaS. He said that SaaS can also be read as "Service as a Software". This means to me that companies today are rethinking their business model to expose their services to be consumed as software. Wow !

    So now you have a choice, you can have no understanding of SOA, you can decide not to get ready to produce or consume these services, no versioning, no exception handling, logging, no registry of anything... It is perfectly your choice.

    I wish Anne's "clients" good luck ... If there is a better dead-end (cul-de-sac) that telling people, no worries, build "services" (how?), you'll worry about everything else later. You know that I know who's going to lose that game.

    Again, to be clear, SOA should not cost that much. If you spent more than 1$ per SOAP Request per Day for your SOA initiative (this is not cumulative, 10 M TX = $10 M spent), you probably overspent. Vendor products are mostly useless to implement SOA best practices. Sometimes they even really get in the way.

    SOA requires that you think hard about what you are doing: about reuse, versioning, about operations & management, about certification, about business models. Maybe that's what really hard about SOA, you have to "think". No easy button, just hard working people that relentlessly get the job done. If you have a brainless management, you listen to silly analysts that recommend changing TLAs every other year, that claim that "this is dead" or "that's what you have to do", no wonder that these same people can only see failures. They are the reason why people fail.
    jdubray