Bruce Richardson of AMR Research predicts the unthinkable: that "rapid adoption of SOA will lead to the end of the ERP market as we know it."
"Here’s the doomsday scenario, circa 2010: SAP and Oracle customers have stopped buying applications from their ERP vendors. Instead, they contract with low-cost Indian or Eastern European integrators to build custom composite apps that sit on top of their ERP backbone."
Why wait until 2010? I've seen it happen in 2005. In one instance, presented in this blog last December, a retailer/distributor was able to forego a major PeopleSoft upgrade by instead putting desired new features out in a service-enabled middleware stack. A recent article in ComputerWorld described how SOA middleware was used to extend a company's SAP-based functions following a major merger.
ERP has long been the application we love to hate. It's huge, expensive, and takes a long time to put in place. The business has to bend to fit the solution, not the other way around. But once it's in, it's too painful to rip out. It should be noted that some folks look at an SOA rollout as being just as bad, or maybe even more painful, than ERP.
So as Richardson suggests, SOA may not overcome ERP with one big swoop, but through a dribbling of features toward the service layer, forestalling or reducing upgrades. And, indeed, much of this work may go to lower-cost offshore development firms and integrators. Ultimately, Richardson argues, it's the integrators who will benefit from the componentization and service-enabling of ERP.
Let's not give SOA all the credit for liberating our enterprises from the monoliths, however. There are other forces at work disrupting the ERP market as we now it. The growing volume of open source offerings in this space, such as Apache OFBiz, SugarCRM, Tiny ERP, and Compiere. Then there are the software-as-a-service offerings, from Salesforce.com for the CRM aspect to Plexus. (Oracle and SAP are already active in the hosted/SaaS space as well.) The Microsoft mass-market mid-market commodization threat also looms, via Dynamics GP/AX.
To beat these new upstarts at their own game, it's a question of how fast the big ERP vendors can break down their applications into digestible -- and presumably lower-cost -- components. SAP and Oracle talk the right talk, but it remains to be seen how far they walk. Richardson observes that SAP and Oracle are investing billions to Web-service-enable their products. But, he adds, "to date, the products are modest. SAP has announced the first 500 enterprise services, while Oracle has yet to release any details on the number it has or plans to make available." Richardson quotes an SAP executive who estimated that his company may have to "provide up to 30,000 Web services in order to fully Web-service the whole suite across all the verticals served." Yikes -- that's going to provide plenty of employment.
Of course, there is another line of thinking that says having a huge, robust ERP system makes SOA unnecessary. As a recent report in CIO noted: "Larger companies whose application infrastructure comes mostly from a single vendor (60 percent or more, according to experts we spoke to) will want to think carefully about whether building their own SOA is necessary or wise." CIOs from both Owens Corning and Whirlpool Corporations state that SAP is their integration strategy -- and they leave all the heavy lifting to SAP.
So which is it -- will SOA kill ERP -- as AMR suggests -- or is SOA unnecessary with a well-integrated ERP? Does SOA take away the pain of ERP, or is SOA even more trying than ERP?