Rick Sherman has a warning for those of you who think SOA will solve all your data integration problems alone: It won't.
Sherman founded Athena IT Solutions, a data warehouse and business intelligence consultancy, but he's perhaps better known for his DM Review column, "The Data Integration Advisor."
In a recent post on Informatica's blog, Sherman asks a good question:
Many companies are building Service Oriented Architecture (SOA) initiatives apart from data integration projects. Since information is one of the key ingredients to creating business value from SOA initiatives, how could this disconnect be occurring?
It's rhetorical, of course, since Sherman goes on to outline exactly why this is happening. The reasons include the fact IT focuses too narrowly on projects, segmenting by technology rather than seeing the whole picture.
Of course, regular readers of this blog shouldn't be surprised. Last spring, we wrote several times about the challenges of data integration and SOA.
He suggests companies start by establishing an Integration Competency Center -- his original post provides links where you can find more help with that.
You'll also need to develop an overall data or information integration architecture, he writes, which should be built using NOT just SOA, but other integration technologies, such as EAI (enterprise application integration); EII (enterprise information integration); ETL (extract, transform and load); and data quality software.
SOA consultant David Linthicum wrote about the problems between SOA and data integration technologies in April, last year. In that article, published on Virtualization.sys-con (be prepared to close the ever-present pop-up ad and hit pause on the always-playing video ad on this site), he noted, "The problem is that many data integration solution are more about information/data than about sharing services, so they're hard fit for many SOAs."
What I'm wondering is how the data integration/SOA conundrum is affected by solutions that publish the data as services.
I interviewed Bill Miller, executive chairman and CTO of XAware, about his product, which made news a few months ago after the company opted to make it open source -- despite a hefty profit from licensing fees. Miller explained that XAware is one of a handful of offerings that make data available as a service, which makes it a perfect go-between for SOA and data integration tools. Other solutions that include this functionality include the TIBCO BusinessWorks products, BEA AquaLogic and Itemfield, which was acquired by Informatica.
Miller said about a third of the company's clients use the product to service-enable data for an SOA or a rich Internet applications environment. Another third use it to build XML structures for business-to-business message exchange, and the final third use this approach to move data in an extract, transform and load-esque application.
Perhaps some enlightened reader will chime in -- otherwise, I'll investigate a bit later and report back.
As a side note: If you missed it, there's been a mini-buzz this week over whether data integration vendor Informatica is in trouble, since all its pure-play peers were acquired last year, and whether it will or should be acquired. Rajan Chandras at Intelligent Enterprise started it with this column, "Is Informatica Losing Relevance?"
Vincent McBurney over at IT Toolbox offered an opposing view Tuesday in his post, "Tackling the myth that Informatica needs to be acquired."
Both are useful reads if your company uses or is considering Informatica.