What CMIS Will Do for Content Integration

Given the level of investment enterprises put into their Enterprise Content Management (ECM) platforms, you'd think they'd take every opportunity to extract maximum value. In a recent InformationWeek Analytics/Intelligent Enterprise Enterprise Content Management survey of 276 business IT professionals, 59% of respondents said their ECM systems could play an expanded role in the enterprise -- if they could be more easily integrated with third-party applications.

That's a big if. Linking apps to an ECM platform is expensive, involving extensive internal development or third-party specialists. That limits the number of apps -- and users -- that can take advantage of the content stored in these platforms.

Enter the Content Management Integration Services (CMIS) standard, a Web services specification that lets competing content management systems share information. Backed by leading vendors, including Alfresco, EMC, IBM, Microsoft, Open Text, Oracle and SAP, CMIS aims to pry open proprietary ECM repositories, thus making the content stored in those repositories more available to applications and end users. For instance, enterprise search software or collaboration platforms could use the standard to find and share content. Web 2.0 mashups could be written to pull information from multiple repositories using CMIS as a common interface, instead of creating custom connections.

This article, which is an executive summary of the report "Silo to Gold Mine: What CMIS Can, and Can't, Do for ECM Integration," explores the capabilities CMIS will enable, which capabilities it lacks and what direction it will likely take in subsequent versions. The full report, which is available as a free download, details early vendor CMIS offerings, provides a detailed look at vendor roadmaps on integrating CMIS, and offers charts and more detail on our Enterprise Content Management survey (registration required).

Why We Need to Broaden ECM Use

Traditionally, ECM vendors have built their platforms like walled gardens, with a limited number of applications and users allowed through the gates. Adding applications typically required organizations to build connectors using vendor-provided APIs, or to hire the services arm of the ECM vendor or third-party consultants to create those connections.

On the repository front, organizations are grappling with the expanding cost of storing unstructured data. Enterprises incur both capital expenses to buy storage space for growing volumes of information and operational expenses to manage that additional capacity other files potentially relevant to the case. The more business content that's stored in a well ordered ECM, the more streamlined -- and the less costly -- that process becomes.

On the resource side, ECM platforms hold a potential gold mine of information. If a wide variety of users and business apps could link to these platforms, all kinds of new capabilities open up, from collaboration to business search to composite applications that repurpose data from multiple sources in compelling new ways. Text extraction, text mining and "voice of the customer" applications and technologies, for instance, are being used to detect customer sentiment, up-sell and cross-sell opportunities as well as risks and threats to enterprises.

It's clear that enterprises want these capabilities. So what's stopping them? Our survey respondents say the most challenging aspects of ECM integration are time and cost, but a lack of in-house expertise is also a roadblock. Organizations also struggle with maintaining their existing integrations, and don't have the resources or bandwidth to introduce more. But help is on the way. Integration on the Horizon

ECM integration has a promising future, thanks to CMIS, which is currently moving its way through the ratification process. The CMIS standard provides a common set of Web services that will enable applications to interoperate with competing ECM repositories. Currently, developers have to build content-oriented applications for specific vendor platforms, such as IBM FileNet or EMC Documentum. If developers want the application to work with multiple platforms, they have to code each integration instance, which increases development costs.

As CMIS support comes online, we'll start to see single applications that work with any ECM platform that supports CMIS. This saves on development efforts. In fact, vendors hope CMIS will do for ECM what the structured query language (SQL) did for databases.

"SQL provides interoperability for developers that want to build data-oriented apps to interact with different database repositories," says Ken Bisconti, vice president, products and strategy, IBM Enterprise Content Management. "So CMIS will do this for content management apps."

Others agree. "For the ECM marketplace to get to the next growth wave, it needs a standard," says Razmik Abnous, CTO of the content management and archiving division at EMC. "ECM platforms need applications. Without those apps, the platform is only half the story."

Abnous says end users don't care about what back-end systems they may be accessing. They just want the information they need in the applications they use. As for ECM customers, CMIS has strong potential to make their ECM platforms more relevant, both for business units that already rely on ECM as well as other groups that could mine those repositories.

In addition, for companies that have multiple ECM platforms deployed, CMIS can serve as a bridge between apps and those disparate repositories. And according to our survey, 58% of organizations have two to three ECM platforms in house already, while 18% have four or more.

Abnous says he anticipates that CMIS can spur development of new apps across a variety of verticals. For instance, as the healthcare industry digitizes patient records and stores as digital files more diagnostic information, such as MRI scans and X-rays, the need to manage that content and make it accessible to the right person at the right time becomes critical.

This confluence of applications and content repositories goes a long way toward explaining the makeup of the major backers of CMIS. Initially launched by EMC, IBM and Microsoft, the group now includes Alfresco, Open Text, Oracle and SAP.

At present, CMIS focuses on two major functions: interoperability and data movement. On the interoperability front, a CMIS consumer (for example, an application) can use the specification to communicate with multiple repositories. For instance, if an insurance company has records in platforms from both Open Text and EMC, a claims-review application could be built that can pull information from both repositories and present it in a portal.

CMIS enables data movement from one to the other without the need for a purpose-built connector. The CMIS specification is being managed by the Oasis standards body. CMIS 1.0's public review period ended in late December, and the group is now reviewing comments. The final standard is expected this year. CMIS in the Real World

Because the standard isn't officially ratified, there are few working implementations of the CMIS specification. However, our free, downloadable report, "Silo to Gold Mine: What CMIS Can, and Can't, Do for ECM Integration," provides a rundown on early vendor CMIS offerings and details vendor roadmaps for integrating CMIS with existing and future ECM platforms. In addition, the report includes more detail and charts from our recent InformationWeek Analytics/Intelligent Enterprise Enterprise Content Management survey. To download the report, click here (registration required).

Andrew Conry-Murray is business editor at InformationWeek and Network Computing. He writes about information management and compliance issues. Andrew has been covering information technology for nine years, including security and network and information management.