DIG IT AWARD FINALIST: OPEN DATA
DOD works to bring data sharing into full bloom
- By Karen Epper Hoffman
- Oct 06, 2016
When it comes to sharing mission-critical data quickly, dueling formats are the enemy. So the Defense Department is developing tools and standards to make such sharing easier and more secure.
DOD’s Data Format Description Language, nicknamed Daffodil, is an open-standard modeling language that ensures that data is properly inspected and filtered as it moves securely across classification boundaries as efficiently as possible. Until recently, filtering most non-XML military data formats had to be done with custom-developed solutions -- each of which required time-consuming and expensive testing and certification.
“Daffodil is also really about solving this whole data format problem once and for all,” said Michael Beckerle, principal of engineering at Tresys, the contractor that has been leading development of Daffodil for DOD.
There are dozens of enterprise data-format tools in the marketplace, but they’re either too expensive or not comprehensive enough for a government agency’s breadth of data, he added. Also, many data-format tools would lock an agency into the company’s proprietary technology.
“An open standard and an excellent open-source implementation are the only way to really change things for the better,” Beckerle said.
“These days there’s lots of news about open-source software in government and open datasets and the tension between openness and privacy or security,” he said, adding that DOD’s case is especially interesting because the agency needed a system that is comprehensive and standardized, with the cost and quality benefits of open-source technology. But many of DOD’s data formats are not published and instead are classified or labeled “For Official Use Only.”
This year is an important milestone for Daffodil because, after almost a decade in development, the technology has become mature enough to support reading and understanding non-XML data in a large-scale way. That means a single Daffodil-based inspection engine can review and transform data contents regardless of their format. In turn, that will allow agencies to more easily comply with new data publishing laws and executive orders mandating the sharing of government data.
Karen Epper Hoffman is a freelance writer based in the Seattle area.