SOA=ERP but worse: fair analogy?

SOA=ERP but worse: fair analogy?

Summary: Many CIOs and IT managers are still smarting from the pain of enterprise resource planning (ERP) system rollouts -- multi-million-dollar megaprojects that required tearing up processes across the enterprise in an effort to automate.  Some have been outright disasters, and many others have simply been money pits.

SHARE:

Many CIOs and IT managers are still smarting from the pain of enterprise resource planning (ERP) system rollouts -- multi-million-dollar megaprojects that required tearing up processes across the enterprise in an effort to automate.  Some have been outright disasters, and many others have simply been money pits.

Jeff Schneider of Momentum knows how to cut to the chase -- making his postings very lucid and thought-provoking.  In his latest blog posting, he warns that SOA will bring more pain and suffering than ERP.

Schneider writes that he scared the bejeezus out of a CIO when he pointed out that an enterprise SOA roll-out will be significantly larger than an ERP implementation. "SOA is a complete overhaul impacting how systems are analyzed, designed, built, integrated and managed. And not just some systems - all systems including packaged applications like ERP."

Not a trivial task, he points out. "SOA requires a rethinking of your infrastructure, development methodology, business impact analysis, budgeting process, organizational design... Don't underestimate the value it provides, the competitive advantage one can assume AND the investment it will take!"

Fair enough. But the beauty of SOA is that it can be brought along in an incremental fashion. As older systems are retired, for example, they can be replaced with standardized components made available to the enterprise, perhaps from the CIO's office. You can have a new EDI system supported by standardized Web services functioning right alongside a big honking ERP system running on AS/400 that has not been upgraded.

Schneider says yes, the incremental approach works, but don't expect a lot of payback. "You can take a slow and steady approach to SOA but be prepared to receive gains that are even slower. Remember - this is really about the 'network effect'."

Will slow and steady win the race?  Or are we in for the mother of all megaprojects, with "death marches" and squashed careers ahead?

Topic: Enterprise Software

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

Talkback

2 comments
Log in or register to join the discussion
  • Talking Turkey

    I have to agree with you Joe. Momentum's approach to describing the challenges of implementing SOA appears to be far too monolithic.

    I certainly begin my own approach by saying something very much similar to your comment: " ...the beauty of SOA is that it can be brought along in an incremental fashion... "

    While there can, surely, be no arguement on the fact that stand-alone applications are being replaced by components that (particularly as they have internal structures that closely track their standardised external interfaces) become much easier to handle as interchangeable components - switched to meet the demands of business managers - it is a mistake, IMHO, to equate that with " ...a complete overhaul impacting how systems are analyzed, designed, built, integrated and managed. And not just some systems - all systems... "

    That kind of talk would leave a nasty stain in most people's unmentionables.

    There does have to be a consistent approach, and there is a learning curve in terms of understanding how that approach needs to be different. I always set the first main objective as; Understand what we mean by services, and start breaking down your current infrastructure in terms of services [even though it can sometimes feel like shoe-horning a live eel into a bucket full of live fishing bait] to learn about service agility, and meeting customer's needs through QoS metrics.

    But a CIO does not have to start 'providing services' just because they finished this exercise. It is often enough just that they can see how things will work to be ready for tommorrow.

    In addition Momentum's approach (again, IMHO), appears to have left out one of the most important elements - external service providers. There Net site, which I have just visited, is very nice. Perhaps it is writen to attract a certain audience (apparently: insecure CIOs and ICT Managers), but I found nothing on outsourcing. There are, of course, up sides and down sides to this consequential feature of SOA but it is my belief [having worked in application services] that the commercial benefits of this will, ultimately, become a commercail imperitive.

    Schneider is trying to have it both ways when he says "yes, the incremental approach works, but don't expect a lot of payback. You can take a slow and steady approach to SOA but be prepared to receive gains that are even slower. Remember - this is really about the 'network effect'."

    Exactly right, the network effect is what makes Web Services and, therefore, SOAs work. Now, show me a network effect that arrived overnight...

    That is precisely why a steady approach is far better. SOAs will change as they mature, suppliers across the supply chain will come up with innovative interim solutions [many very highly funtioned and robust solutions already exist] to reduce the impact on your ICT investment cycle, and the outsourcing angle will grow alongside the network effects.

    Although I would normally caution that it is an extremely bad idea to outsource a problem, another feature of evolving SOAs (and the associated 'biosphere', or 'support' network of suppliers) will be that actually switching some business processes to SOAs later, through outsourcing, will actually help to overcome existing problems. That synopsis assumes you can wait, of corse.

    In summary then: Humbug!
    Stephen Wheeler
  • Yes, but when will SOA be exposed via ERP??

    From my perspective, as an ERP vendor working on SOA right now, it's highly likely that SOA will be used under-the-hood for a few generations before it's fully exposed externally. There are huge advantages to code reuse, design by BPM and model driven architectures, but the toolsets and shared experiences such that a customer could readily and safely rejig the components of an ERP are currently very limited. Also, while web services provide a model for SOA, there are still many issues around security standards and QoS that preclude exposing the bulk of an ERP as SOA components and expecting it to hang together.

    Our product roadmap, for example, is to rollout external-SOA in a limited fashion at a very coarse grained level initially, then refine the 'grain' as suitable orchestration tools, security models and support methodologies are standardised. Of course, there will a much higher level of components within the system, but they will not be exposed for in the first iteration.

    So, I think the view that "SOA=ERP but worse" very much overstates what will happen by presuming that ERP vendors can't separate marketing hype from operational reality.
    mpanosh