Communities of Data

As it works to develop a data strategy geared towards making data visible,
accessible, understandable, interoperable and trusted, the Department
of Defense is putting heavy emphasis on forming communities of interest.

By Peter A. Buxbaum
 

As it works to develop a data strategy geared towards making data visible, accessible, understandable, interoperable and trusted, the Department of Defense is putting heavy emphasis on forming communities of interest (COIs).

The premise behind network-centric operations, the Global Information Grid and service-oriented architectures is that information throughout the enterprise can be deployed wherever it is needed. DoD’s approach of developing a SOA means offering data and modular capabilities as services across the enterprise and decoupling data from applications. The major sticking point to the realization of these ideas is that data from one area of the enterprise is not necessarily understandable elsewhere.

While a data strategy is essential to achieving the vision of net-centricity, that still leaves the question of how to achieve these data goals. One possibility would be to impose top-down, DoD-wide data element standardization. That approach was rejected, however, in favor of forming COIs, which are groups of users who exchange information in pursuit of shared goals, missions or processes and who therefore must have a shared vocabulary for the information they exchange.

“It would have been a monumental project if DoD had decided to start at the top,” said Richard Hull, chief scientist at Modus Operandi. “They would have had to get everyone to agree what to call every object in the DoD domain. How long would that have taken?”

Establishing communities of interest provides for a bottom-up approach driven by people that have knowledge of the domain they are dealing with in the COI. “Data harmonization is much more sustainable in those circumstances,” said Hull. “People close to the data are the ones figuring out how to make it discoverable. They are in a position to say, ‘We’ve got this data and we expect that this data could be reused outside our organization.’ So they publish it as a service in such a way that other folks can get access to it and make use of it.”

“Different organizations have different ways of describing information,” explained Leslie Winters, chief of the joint data strategy division of U.S. Joint Forces Command. “In the command and control area, we have identified 500 sources of authoritative data and 40 different vocabularies that are used. By developing a shared vocabulary, communities of interest are able to describe the same thing in the same way. If vocabularies are not harmonized you can potentially have a real mess down the road.”

USJFCOM heads the data harmonization effort within DoD’s command and control portfolio.

“The biggest challenge is that commanders want to have situational awareness so that they can see and understand activities on the battlefield and anticipate enemy actions,” said Stuart Whitehead, USJFCOM’s executive director for joint capability development. “This requires access to information from a variety of sources depending on the type of mission involved. The information a commander needs will also change with time and circumstances. This requires better access to timely and accurate information and putting those information elements together in a way that is understandable.”

Shared Vocabulary

Developing communities of interest involves bringing together different defense agencies that all deal with the same function or process, noted Michael Krieger, director of information management in the office of the DoD chief information officer. Communities of interest can be of wide applicability, such as logistics or meteorology, or represent such discreet areas as strike warfare, time-sensitive targeting and geospatial capabilities.

The shared vocabulary that results from the work of the communities of interest takes the form of metadata extensions to the core set of metadata set forth in the Defense Discovery Metadata Standard (DDMS). DDMS is a specification that, along with various eXtensible Markup Language (XML) schema, is being implemented throughout DoD in order to tag electronic resource holdings.

Data tags represent the content of the information resource and facilitate an efficient and efficacious search for the information. Instead of searching through an entire document, only the metadata is searched. This results in a more efficient search and yields a higher fidelity result.

DoD Directive 8320.2 mandates that data harmonization and standardization be accomplished through the establishment of communities of interest. But that doesn’t mean that there is not some top-down persuasion involved to actually make that happen.

“The problem with trying to tackle data issues is getting the owners of capabilities within the department to agree to use communities of interest to tackle problems,” said Krieger. “I don’t want to be the person directing everyone to do this. I’m more about pointing out communities of interest as a useful approach to tackle this kind of problem and engaging program managers in a discussion. It’s a classic problem of defense transformation: getting people to adopt this approach and directing a change in organizational culture.”

For Krieger, participation in communities of interest and the success of their endeavors are driven by the perception of a significant problem that crosses organizational lines. “The community of people who own or will be making use of the data need to participate in order to work the semantic translation,” he said. “The problem has to be painful enough before people are motivated to join a community of interest addressed to a joint problem. The bottom line is that whoever is a stakeholder in the problem needs to play so that all involved can adopt a solution together.”

Krieger points to “a couple of big wins” in DoD’s community of interest data strategy: the maritime domains awareness and the joint strike communities of interest. “These demonstrate that this is an approach that works,” he said. “Maritime domain awareness delivered an initial capability in nine months.”

The maritime domain awareness community of interest, which extends beyond DoD to include the Department of Homeland Security, and Department of Transportation and other federal agencies, is currently working on its third spiral of deliverables.

Joint Strike

The joint strike community of interest is working on its second spiral and has already standardized rich semantic elements enabling the sharing of weather, air defense and other information. The joint strike community of interest is also notable for having delivered the “cursor on target” semantic capability.

The cursor on target project developed a joint data model, schema and vocabulary for expressing the what, when and where of a potential target. Key to the success of this project was an agreement among the services to adapt the best attributes of the Air Force’s existing solution. The emergence of commercial Geography Markup Language (GML) as a standard for representing geographic features and the adoption of GML within DoD was also a key development driving the success of the cursor on target project.

Another community of interest, centered at the 45th Space Wing at Patrick Air Force Base, Fla., has been focused on streamlining data sources for DoD’s space community. “A study showed that the Air Force IT infrastructure included 5,000 databases,” said Modus Operandi’s Hull, who works with the COI, “and estimated that only 20 percent were authoritative data sources, while the other databases were basically copies of the original authoritative data that was being used for other purposes.”

As part of the Air Force’s network-centric strategy, the secretary of the Air Force ordered that users of Air Force data be pointed to the original authoritative sources.

The space launch COI, managed by the 45th Space Wing, is designed to develop a service whereby any DoD unit seeking to launch a satellite would be able to search for and discover launch schedules at Cape Canaveral or elsewhere.

“This is the kind of information that you would otherwise get by calling around on the telephone, a very manual process,” said Hull. “The idea of the community of service was to develop a Web service through which you could quickly find the information as long as you had the authority to pull it down.”

The COI is involved in a two-year project to harmonize the metadata surrounding satellite launch terminology. “Some databases may call it launch schedules, others may call it launch operations,” he said. “You need something to help translate a query in order to find the relevant information. We found this process to be more challenging than we expected.”

Data harmonization cannot be accomplished strictly with a bottom-up approach, as Krieger suggested. At USJFCOM, DoD’s command and control communities are getting a nudge from the top.

“We have to take a portfolio approach to command and control,” explained Winters. “There are several C2-related communities of interest that are functioning well with a bottom-up approach that is geared to their specific mission areas. That is fine, but we are trying to influence data strategy for C2 with our counterparts across DoD on a different level.”

“Communities of interest are oriented to a particular functionality,” added Whitehead. “But command and control touches many COIs and many data sources that may have no relationship to a COI. We take the broadest view of command and control to view what is common in the data across the entire portfolio. That way different participants working on different issues can all come together in a way that promotes commonality across the portfolio.”

What JFCOM’s joint data strategy division has done is to develop and release a cross-service command and control semantic core. “The point is to establish common data element representations across C2,” Winters explained. “Communities of interest can use this as a starting point as they develop mission-specific solutions.”

JFCOM has also recently released a conceptual data model, a reference schema and a core metadata registry for command and control.

Harmonization Process

JFCOM’s efforts are meant to ease the burden of communities of interest in the command and control area by jumpstarting the data harmonization process. But the heavy lifting must still be borne by the COIs.

What makes for a successful COI, according to John Scott, a director with Mercury Federal Systems, is the commitment made to each other by the people who participate in it. “When the people in a community of interest agree to data specifications, they are more likely to stick with it than if organizational higher ups impose it. They want to keep these people as colleagues and friends. There is a social contract component to communities of interest. That is why they are so powerful when they are run correctly.”

But that doesn’t mean that everyone is enamored of COIs. “We’re getting some resistance to communities of interest,” said Krieger. “We view this as a good sign, because when you are involved in a transformational process, resistance shows that something has hit home.”

Where Krieger sees resistance is primarily from the owners of tightly coupled systems. “What we are asking them to do is to break apart those applications, to decouple data from applications, and to deliver incremental capabilities,” he explained. “This is a very different approach than what they have been used to.”

The biggest impediment Krieger must overcome is the fear among program managers that the new approach poses a threat to them. “I tell them that communities of interest working in concert with the acquisitions system is a good tactical risk-mitigation approach across programs of record,” he said.

“We are trying to deliver 80 percent solutions quickly and to get those capabilities out to the field. We don’t want to wait for two or three years while a project develops. Communities of interest are helping us spiral up capabilities quickly.

“Beyond that,” Krieger concluded, “we are trying to identify some critical joint problems and to establish a dialog on how the approaches taken by communities of interest are able to address those problems.” ♦