A new 12-rack capacity computer room was built in NAA's Mitchell building in Canberra for the project and came online in March this year. The room and hardware cost about $1.2 million (US$880,000). NAA will look at upgrading storage for this facility within six months.

The application software, developed in-house, was open sourced by the NAA: Xena, written in Java, which converts electronic records into a standard XML format, Digital Preservation Recorder to track the actions of record, and a Quest database for keeping information about objects in a repository.

Pearson said the project has become one of digital preservation.

"Traditionally in the paper world, we take about 5 percent of paper from [government departments and] agencies, but how much we will take in digital we don't know, it may be 5 or 10 percent," Pearson said, adding that since no agency tells the NAA how much digital data they'll submit, it makes capacity planning "quite frustrating".

McKinnon said the digital archive will eventually get to petabyte scale. Agencies have to recognize this, she said, and cited as an example one unnamed agency which has 35TB of e-mail alone.