Open platform preferred for digital archives

An ERA advisory committee recommends open architecture standards if security constraints allow.

The National Archives and Records Administration should consider building its electronic archives using open architecture standards, according to members of an advisory committee assigned to confer with NARA.

The system’s security requirements, however, will dictate whether the Electronic Records Archives could have an open architecture, the committee said last month.

ERA is a $308 million project to preserve government records regardless of format and make them accessible on future hardware and software.

The 18-member ERA Advisory Committee will consult with archives staff on issues related to the system’s purpose, technical development and use. Internet pioneer Robert Kahn, who is chairman, chief executive officer and president of the Corporation for National Research Initiatives, serves as the committee’s chairman.

“He’s a very strong advocate of a model of open architecture,” said Ken Thibodeau, ERA program director at NARA. Kahn urged the advisory committee to develop an open architecture for the ERA, and committee members concurred.

An open architecture would allow the system’s components to be interchangeable. Engineers could replace components to adapt to future changes in technology. The committee said an open architecture would be important in helping manage the evolution of the system.

The committee also agreed to form two subcommittees, one focused on architecture and the other on usability.

Experts on the architecture subcommittee come from diverse fields, including information technology and archival science. Both subcommittees have experts who are knowledgeable about IT, archival science, history, genealogy and education. Kahn serves on the architecture subcommittee.

Security issues

Thibodeau stressed that security policies will ultimately determine NARA’s ability to create an open architecture and implement open-source software. “While there will be a single design for ERA, it will be implemented as three separate systems encompassing unclassified, secret and top-secret information,” Thibodeau said. “We want the same technological baseline in all three systems so that we are all managing one technology. We will have to get approval for introducing any open-source software.”

The National Security Agency is responsible for crafting a policy regarding security for systems that handle classified information.

Open-source software, if approved, could be useful in overcoming the challenge of obsolete technologies. For instance, if someone in the future develops nonproprietary software that could convert an obsolete spreadsheet into an Extensible Markup Language spreadsheet with the potential to be accessible for more than 20 years, NARA would want ERA to incorporate that software.

“What we want to avoid above all things is developing a lot of software that’s just coded for NARA’s system,” Thibodeau said. The agency would prefer to take advantage of commercial and open-source software that is readily available, he said.

Providing public access to the system will also require technological flexibility and interoperability. Committee members recommended that NARA let companies like Lexis-Nexis repackage some ERA content so that legal and academic researchers can access it.

Most major universities have digital libraries designed to tap into a wealth of research databases. NARA wants ERA to be interoperable with those academic systems. Eventually, NARA also wants to share ERA software with state digital archives.

Avoid system failure

NARA awarded Lockheed Martin a $308 million contract to design the system. The agency is counting on the experience of Kahn and the other advisers to ensure that the project does not become another one of the government’s major systems failures.

Without having seen Lockheed Martin’s design in detail, however, the committee said it is not ready to give any formal advice. “We decided that we would pursue an open architecture approach…but whether that’s going to be possible given some of the security constraints, it’s a little difficult to know,” Kahn said, adding that “we still have to see what Lockheed has proposed.”

Mining archives data

As some advisory committee members discuss the architecture of the National Archives and Records Administration’s Electronic Records Archives (ERA), other members are focusing on how ERA’s broad array of patrons will access it.

David Carmicheal, president of the Council of State Archivists and a member of the ERA Advisory Committee’s usability subcommittee, said the concentration of records in one databank could help users discover resources they would never find on the shelves of a library archive.

Traditionally, archivists have waded through collections for researchers. With the new digital archives, lawyers, students, government officials and others could search collections on their own, said Carmicheal, who is also director of the Georgia Archives.

“The concept in my mind is that there would be the normal public interface but that groups of people with expertise in a certain domain might be able to mine the data more deeply because they know what they need to get out of it,” Carmicheal said. “It is impossible for the archivists to anticipate every need.”

He said much about ERA is still unknown. “What we want to do is sketch out the ideal and then work with realities,” he said. “We may come back and say everybody should be able to mine this data. The government may say [that] for security reasons not everybody should be able to mine this data.”

— Aliya Sternstein