New hope for data exchanges
- By Joab Jackson
- Jun 24, 2005
GCN Photo by Olivier Douliery
With the revision of the Federal Enterprise Architecture's Data Reference Model, expected in December, federal agencies could finally have solid metrics to evaluate how well they share data, said Michael Daconta, metadata program manager for the Homeland Security Department.
Such metrics could break the barriers that have kept agencies from sharing information, he said. Daconta heads up the working group in charge of revising the DRM, the first draft of which was released earlier this month.
'This is a detailed blueprint of how organizations are going to describe the structure, categorization and exchange of their information,' Daconta said at the Data Reference Model Public Forum, held recently in Washington in conjunction with the Federal CIO Council's quarterly Emerging Technology Components conference.Lacking details
When Office of Management and Budget issued the first version of the DRM last fall, Daconta criticized the specification for not sufficiently detailing how agencies should represent their data. He went on to lead a working group to revise the specification.
The working group is now soliciting feedback from government agencies before it submits its revised DRM to OMB this fall. The Federal CIO Council's Architecture and Infrastructure Committee, OMB's Federal Enterprise Architecture Program Management Office and DHS support the working group.
The release of the DRM draft is an important milestone in the federal government's efforts to better share information, said OMB chief architect Richard Burk, who also spoke at the conference.
The last of the five components in OMB's Federal Enterprise Architecture to be released, the DRM is, in many ways, the linchpin that holds them together.
OMB strongly encourages agencies to use these reference models to define their enterprise architectures, which will facilitate cross-agency analysis and investment.
Each reference model addresses a different aspect of the enterprise architecture: The business reference model describes the services the agency offers; the technical reference model describes the technology supporting the missions; the performance reference model describes the metrics to be used, and so on.
The DRM 'supports each of the [other] reference models,' said Mary McCaffery, the Environmental Protection Agency representative of the working group. It provides the structure that 'facilitates the development and effective carrying of government data across communities of practice and lines of business,' McCaffery said. Agencies can use the DRM schema to describe their data, specifying what format the data is in, what topics the data addresses and how the data can be accessed.
When OMB first issued the DRM last fall, Daconta noted that the specification was fuzzy in some areas, which would lead agencies to question how it could be implemented. For instance, it did not adequately separate structured from unstructured data, nor address matters of security and privacy.
By coding the DRM in Extensible Markup Language, the DRM could be used to measure how well an agency is prepared to share information, Daconta said. In other words, the degree to which an agency uses the DRM could provide a direct measurement of how well it is meeting a requirement to share information.
Lawmakers and the public have be- moaned the continued lack of information sharing among agencies, which has improved little even after the events of Sept. 11, 2001, have shown the detriments of not sharing information.Data sharing
'If, God forbid, another event occurs, we should be able to say 'Did you have the information? Did you do your DRM?' If not, heads should roll,' Daconta said.
An agency that develops a system for use in-house would not have to write a DRM description of such a system. 'This is not a drive to harmonize all of your information. That is too big. We are talking about the information you either currently share or will share within a year,' Daconta said.
By publishing the DRM, the public can find the data through a taxonomy available on a Web site. Other agencies could discover resources through some sort of federal central registry of DRMs describing the resources available.
With the release of this draft, the working group is now soliciting feedback from government agencies on how to improve the model. The group has set up a Web site and public mailing list to receive comments. To get to the site, go to www.gcn.com and enter 433 in the GCN.com/search box.
It will also hold open workshops July 19, Aug. 16 and Sept. 23. The group will add new features to the document through Sept. 14 and will submit the finished draft to OMB Nov. 17. OMB should publish the revised DRM by Dec. 17, according to Daconta.
In addition to revising the schema, the working group will also draft documentation that describes the management process that agencies should undertake to collect and register the data. In the future, the working group will draft specifications for how to harmonize multiple DRM-based data models. The DRM allows agencies to translate an existing data model to the DRM, as long as they used some other widely used open standard to describe its model.
'No one has tried this before at this scale,' Burk said. 'This will provide an open and well-documented standard that will enable the organization and categorization of government information in the ways that are searchable and interoperable across agencies.'