Federated repository management--making CM play well with others

As Emery explains, "Most larger user organizations have multiple repositories--so the idea of enterprise content management is more of a marketing term than a reality … especially with mergers and acquisitions and simply by the nature of the way people buy software, which is based mostly on budget, not architecture ... So it makes it impossible for one big company to have one document repository." The fact is, customers have had enough, she says, and some CM vendors are having a hard time getting users to buy without having FRM ability.

The problem, however, now for vendors is not being the only CM system in an organization--it's being the anchor CM infrastructure upon which all the others depend and which has the almost exclusive opportunity to proliferate. So the notion that one vendor is really going to use another's enabling software to divest itself of controlling data for a big customer is a bit ludicrous, says Medina.

The two FRM methods
Nonetheless, every major CM vendor except Hummingbird has an FRM solution on the street or in the works. Most use one of two methods to accomplish FRM. The dominant one is what's known as the "Venetica bridge." IBM bought FRM specialist Venetica more than a year ago to give its CM-related products that capability. IBM renamed it WebSphere Information Integrator Content Edition (IICE) and allied with FileNet and Open Text to employ that "bought" solution to quickly give customers FRM. Other CM vendors have or are "building" custom solutions. Vignette's Virtual Repository, for instance, is a custom FRM product that invests the Vignette ECM Suite with FRM capability.

According to Toby Bell, research director, High Performance Workplace, Gartner, the Venetica bridge is at best a stopgap solution whose strength is that it can be deployed now. It's likely all CM vendors will migrate to the Vignette approach eventually. "The right way to do [FRM]," Bell explains, "is to extract metadata from all those silos and manage the metadata as a separate layer--and administration similarly--so administrative control of access and security rights and profiling, etc., associated with the independent repositories is consolidated and centralized." FileNet and Documentum are headed in that direction with their proprietary FRM projects. Bell says the strategy enables better compliance, life cycle control and processibility in terms of claims handling and the like, as well as for user context tied to their profiles, subscriptions, rights and so on.

In contrast, he says, "Venetica focuses on the relationship of one content repository to another," with no master overarching layer for administrative control, metadata markup and the like. Bell explains, "This is not an enterprisewide approach to mapping metadata--there's no global taxonomy … that allows you to add an abstracted layer to better define and revise existing content."

Emery elaborates, saying that Venetica is "more of a one-to-one setup instead of a generic search across all repositories--the bridge actually creates a kind of a gateway into specific applications like CM, DM [document management] and RM [records management] and their repositories." As you can imagine, that approach means prepping the legacy architecture of repositories for integration, fusing them and managing them with a master administrative metadata map. That is the IICE approach--integration- and service-intensive.

Vignette, on the other hand, can sell you a single repository that can handle all types of structured and unstructured data like COLD, images and documents, and works right out of the box. The bad news is that to get full-bore FRM, you need to buy its records and documents module to perform the federation, its portal as the content interface, Web Content Management to leverage scattered content in Web applications for activities like self-service via a Web site, and Vignette Business Integration Studio to do the advanced integration needed to use and deliver, for example, structured content in various non-CM apps, as well as unstructured content in CM systems. The good news is for simple federation, the records and documents module is sufficient and boasts pseudo plug-and-playability--and with those related products Vignette offers a functionality ladder the customer can scale as they need more complex FRM.

The upside to the Venetica approach, on the other hand, is that IBM claims to have adaptors for at least Open Text and FileNet systems and can customize ones for other CM systems. Medina, for one, though, is not convinced IBM can do advanced FRM beyond search and retrieval like revision and records management. He explains that the downside to the Vignette approach, by comparison, is that most of its out-of-the box adaptors are not for other CM systems, but rather for popular enterprise apps like ERP (enterprise resource planning) and messaging (though it can customize CM ones). But then it can also perform advanced FRM--various actions on a document throughout its life cycle, from search through RM.

Medina says that FileNet meanwhile is developing its own FRM code that it will offer in addition to the Venetica bridge, and Documentum can do FRM now via Content Integration Services, an integration-intensive approach that is not yet productized and has no FRM module related to it, though Documentum promises a product by Q1 ‘06. As yet, Stellent, Hummingbird and Interwoven (interwoven.com) are still undeclared citizens of the nascent FRM nation. They may never productize FRM but offer it as an integration service focused on the specific platform, application and market requirements of their installed base.

Bell says Hummingbird, for instance, is the only .NET-based CM vendor and offers complex collaborative and document management but is weak in imaging, workflow and financial services where most of its J2EE-based counterparts like FileNet and Documentum excel. So an FRM service from Hummingbird will be .NET-based and pitched toward collaboration-intensive organizations in legal markets, while IBM and FileNet products will go after the imaging/workflow J2EE-based customers in finance and insurance, EMC Documentum after its J2EE-based pharmaceuticals-intensive captive customers, and Open Text after its J2EE-based ones whose primary requirement is usually advanced collaboration.

The big boys
As with other CM capabilities, the CM pure-plays are being encroached on in FRM by the big infrastructure vendors, IBM, Oracle and Microsoft. Bell thinks the first two, being massive database players who've mastered structured content, aspire to extend their control to unstructured content (which comprises 80% or more of content in apps like imaging, workflow and document management) by migrating it from legacy systems into their relational database infrastructures with a consolidation, instead of a federation, strategy.

For instance, Bell says, Oracle's recent acquisition of FRM vendor Context Media (contextmedia.com) lets it pull off bi-directional integration with CM systems and agglomerate data into Oracle 10g via its Content Services content integration product. "Microsoft," on the other hand, he says, "can do both consolidation and federation since it owns the content creation toolset as well as the routing/approval mechanisms with new workflow capabilities in Outlook and Exchange--so they can already provision content well."

Bell adds that Microsoft SharePoint now competes with IBM WebSphere "as the user's daily habit outside of e-mail, and [the two products'] relative content stacks involve everything from storage up through context [access to data and apps through the portal interface]." Bell even believes that SharePoint's proliferation as probably the de facto, "affordable" portal and soon-to-be CM solution will require any CM FRM product to interoperate with it so users can move content between their legacy CM and other apps and SharePoint.

But let us not forget SAP. Bell contends that "SAP is the heartbeat of a lot of businesses" and all CM players want to be the default repository or distribution mechanism for data in SAP. Bell's guess is that SAP will want to control it as well, and will either buy or build more content handling capabilities, including FRM.

Critical differentiator for CM pure-plays
What the big guys lack will be what the CM pure-plays will have to rely on to differentiate themselves. Bell thinks that is "a level of process to aid in the creation of content-enabled vertical applications, which are really what makes ECM valuable for the business buyer who has replaced the technology buyer in terms of managing the spending and whose focus is on ROI." So in his view, CM pure-plays must focus on federation of content within existing heterogeneous CM apps and sell themselves as delivering better ROI for transforming unstructured content into semi-structured data. It will be some time before Microsoft or the database giants can compete with them on those terms.