In-Depth
With Project Unity, SAS and DataFlux Walk a Fine Line
Project Unity is an ambitious -- and especially fraught -- enterprise for SAS and its DataFlux subsidiary.
Last month, SAS Institute Inc. and its DataFlux data quality (DQ) subsidiary announced a reorganization: a shift of SAS' data integration (DI) assets -- including SAS Data Integration -- to the DataFlux side of the business. The impetus, SAS says, is unity -- Project Unity, to be precise. That's the name for the combined SAS/DataFlux effort. Its goal, according to SAS DI pointman Ken Hausman, is to develop an all-in-one platform for DI, master data management (MDM), DQ, data federation, text analytics, and just about any other kind of integration scenario customers can imagine. The move could be tricky for SAS, however.
For one thing, there's SAS Data Integration, which has become a big underground sensation of sorts in the DI segment. Industry watchers typically place SAS DI among the market leaders, along with stalwarts such as Informatica Corp., IBM Corp. (with both Ascential and DataMirror), Oracle Corp. (with Oracle Warehouse Builder and the former Sunopsis ETL technology), and SAP AG (with Business Objects Data Integrator, nee Acta).
In the near term, SAS Data Integration isn't going to disappear. In spite of its having moved (development team and all) over to DataFlux, Hausman says that no decision has yet been made about its branding. There's a sense, however, in which SAS must walk a fine line, ever respectful of the separate SAS Data Integration and DataFlux brands. It must be attentive to the road map and goals of Project Unity (which emphasizes improved interoperability and synergy between SAS Data Integration and the DataFlux suite). Furthermore, SAS must be careful that the coupling of SAS DI and DataFlux isn't so tight as to make them dependent on one another for common DI or DQ scenarios. Customers accustomed to deploying SAS DI largely for its ETL facility, or DataFlux for its best-of-breed DQ expertise, would likely balk at a bulkier package with an as-yet-undetermined price tag.
The situation is particularly fraught on DataFlux. SAS acquired DataFlux more than seven years ago. Since then, DataFlux has carved out an enviable position as a quasi-independent purveyor of best-of-breed data quality tools. It's a good position to occupy: over the same period, several DQ best-of-breeds -- Group 1 Software, Similarity Systems, and Firstlogic, to name a few -- have been acquired by Pitney-Bowes, Informatica, and Business Objects, respectively.
After seven years, DataFlux retains a patina of independence. That's no accident. SAS and DataFlux officials have been careful not to play up the relationship between the two. They've been especially anxious to avoid an appearance of untoward closeness -- e.g., of SAS borrowing (too) liberally from DataFlux or vice versa. In short, they've tried to avoid arrangements that might call into question DataFlux's independence. (In real-time connectivity scenarios, for example, SAS either refers customers to DataFlux or brings DataFlux into the mix. SAS officials stick doggedly to this position, demurring when asked about the potential benefits of simply incorporating DataFlux' real-time technology into SAS DI. (See http://esj.com/Business_Intelligence/article.aspx?EditorialsID=8875.)
TDWI Research’s position (stated in its recent Technology Market Report on Data Integration Tools) is that the DI tool market is progressing from suites of disconnected DI tools toward platforms of fully integrated tools.
For example, for several years now, a number of data integration (DI) vendors -- namely Business Objects, IBM, Informatica, Microsoft, Oracle, and SAS/DataFlux -- have had suites of tools or modules, one tool each for individual DI and DI-related data management practices. These practices constitute a long list, including ETL, data federation, data quality, data profiling and monitoring, stewardship, text analytics, and master and metadata management. The main benefit of a tool suite is that users have a single vendor to work with for the acquisition and support of multiple, related tools. The problem, however, is that tools and modules of the suite (especially when the vendor acquired them, instead of building them) do not integrate and interoperate deeply, if at all.
The DI platform seeks to correct the suite’s lack of interoperability via the full integration of common tool elements, including those for deployment (metadata, servers, security, interfaces for data access or interoperability with other tools) and development (user interfaces for design, collaboration, management, security). Ultimately, this is what users want, because they’re progressively practicing multiple DI practices together in a single, large project, ideally deployed on a single united DI infrastructure. The organizational changes and product improvements of Project Unity should propel DataFlux down the road toward a truly unified data integration platform.
If nothing else, Project Unity raises questions about the branding and go-to-market strategies of SAS's and DataFlux's traditionally separate DI and DQ toolsets. Hausman says there is no (essential) conflict; Project Unity is as much about realignment as it is an effort to create a best-in-class integration platform, he explains.
"Part of the strategy is about how the two companies play in the market. SAS is really better known as provider of business solutions that take advantage of our platform and probably more specifically our strong analytics focus. DataFlux has been much better known as a data quality provider, which is more on the IT side than on the business side. When you think about it, SAS Data Integration is basically on that [IT] side, too, so I think that's one of the directions that we want to continue to foster," he comments.
"[DataFlux is] more well known for selling to IT and SAS is well known selling [its] solutions to the business side. Project Unity will be a more comprehensive, integrated, full-function data integration application that includes data quality, that includes real-time integration."
How will SAS and DataFlux market Project Unity? Will SAS Data Integration eventually go away?
No one at SAS has yet made any decision about what to do, Hausman stresses, despite its shift to DataFlux. "The team from SAS who were designing our Data Integration Studio and some of our other DI folks have been transferred to DataFlux. People from testing, not just the R&D folks, but folks who focus on the user interface design -- these people don't report to SAS anymore. They now officially report to DataFlux. DataFlux has taken over the design of this next-generation application," he explains.
As Hausman points out, "DataFlux is better positioned to carry the torch forward when it comes to a focus on IT and technology and tools, so they're carrying the banner of data integration as part of Project Unity. There has not been a decision as far as I know about branding," he says.
Pricing is unknown. Will SAS Data Integration or DataFlux DQ customers be able to purchase and license these products as they've always done? At this point, Hausman concedes, neither SAS nor DataFlux has made any decisions in this regard.
Nor have any decisions been made as to whether (or if) Project Unity will be a highly modular (i.e., a la carte) offering, where customers can pick and choose which features they want to use or discard. "I'm under the impression that if you buy it, you get it all. There may be add-ons, like we had with the current DataFlux product … where you might not need to buy everything, but at the same time, the focus is to provide that single GUI that allows the user to use the particular technologies that really make sense for whatever initiative they're working on," he indicates.
"I can't tell you what it's going to look like, but I can tell you that they have designers working on it to make sure that takes all of the best features of both products, the latest technologies to make it easy to use."
At this point, Project Unity really describes a data integration category. Instead of maintaining separate DI and DQ toolsets (on separate sides of the aisle, no less) -- and as a consequence of the fact that DI and DQ need to interoperate with one another in a growing number of integration scenarios -- Project Unity will bring them both together under one proverbial hood, complete with a single GUI and a consistent look and feel.
"Today, SAS has a pretty full-featured set of DI services as part of its Enterprise Data Integration Server, but integration between DataFlux and SAS hasn't been overly deep. It's been strong, but it's only been that way in the sense that they're two completely different applications that are integrated and in some ways designed to work together. They are still two separate applications," Hausman explains. "With Project Unity, there won't even be the concept of compatibility. It's just going to be one application. You run it, you turn it on, you have data quality tools, you have data integration tools."
Hausman assured us Project Unity is taking shape. Initial delivery is pegged for late 2009, although -- once again -- he demurs when asked to talk more specifically about what a finished product will entail.
"I honestly don't know how the product will be configured and what options it will have. It's still in the architectural design phases. They pretty much know the broad outline [of what it will look like]," he says.