Information Security for Digital Assets at MATRIX

MATRIX runs its operations on several servers kept in a climate controlled, physically secured room; these servers run the Debian distribution of Linux. Approximately 2.7 TB of data is stored on the servers as of July 2009. This includes the H-Net e-mail lists, book reviews, and other data; digital collections based on KORA and other repository software; internal working user files, including e-mail files; web data; and software programs such as the KORA repository system, LISTSERV list management system, and others. Backup and storage procedures for audio and video files stored on removable media will be addressed separately.

Data Stored on Servers

Incremental tape backups of the data stored on the MATRIX servers are performed daily using the NetVault backup software application, with a full backup performed on a weekly basis. Tapes are taken to the Cyclotron building across the Michigan State University campus and exchanged for the tapes stored there the previous week. Backup tapes cycle through the system approximately every six weeks and are replaced as needed, such as when a cartridge breaks. The MATRIX systems administrator keeps a wiki-based log of all tape backups. A Redundant Array of Inexpensive Disks (RAIDs) divides and replicates data among multiple hard disk drives for better system reliability.

MATRIX has also adopted two offsite storage plans to better protect and ensure the continued availability of its server-based data. First, an additional full set of backup tapes is created every four months and logged into the MATRIX tape backup wiki. Through an arrangement with the Michigan State University Archives, MATRIX plans to store these tapes in a secure, climate-controlled storage facility in nearby Lansing, Michigan. Until this arrangement is in place later in 2009, the tapes will continue to be stored in a cabinet in a climate-controlled, minimally secured office on the MATRIX premises. These long-term backup tapes are kept on a three-year retention schedule.

In addition to maintaining the long-term backup tapes at the Lansing storage facility, MATRIX has established a reciprocal storage arrangement with the Inter-University Consortium for Political and Social Research (ICPSR) at the University of Michigan, Ann Arbor, 60 miles from the Michigan State campus. On a daily basis, MATRIX uses rsync software to synchronize and copy its data into “dark” storage at ICPSR—that is, storage that cannot be accessed by general users—and MATRIX provides the same service for ICPSR data.

All of the backup and redundancy systems described above will be helpful in disaster recovery. The RAIDs will provide redundancy in the event of a local hardware failure. If data on the server is damaged or compromised, it can be recovered from the full backup tapes stored at the Cyclotron. In case of concern regarding whether “bad” data was backed up, uncompromised data may be recovered from the long-term backup tapes. If the equipment on the MATRIX premises and the tapes stored at the Cyclotron are compromised, data may be recovered from the long-term backup tapes stored offsite or from the copy of MATRIX data in dark storage at ICPSR.