From the Executive Director: Follow-Up on August’s Service Interruption

In late August, HathiTrust experienced a multi-day service interruption stemming from a network outage at the University of Michigan. U-M Security purposely separated the U-M network from the internet to protect against threats to the university’s network.  Because our infrastructure resides in a campus data center at the University, this action also served to protect HathiTrust’s systems and data.

Since 2010 we have mirrored our preservation and access infrastructure at the University of Michigan and Indiana University, which helps to ensure that we can meet high user demands and remain online in the event of a problem at one of our sites.   But the network outage at Michigan was so significant and so prolonged that some functionality was unavailable even while many other services remained online.  Unfortunately this occurred at the start of the academic year for many of our members, and we apologize for the confusion and disruption to access to HathiTrust.  Within a few days, the University restored the network connection and we resumed normal operations almost immediately. Thank you for your patience during this time.

We have since begun internal reviews to identify ways we can improve our response and improve service continuity in the future. We, and our members, consider integrity and reliability of systems and data on a regular basis. In a Scholarly Kitchen article Roger Schonfeld used the security incident and network outage at U-M to reflect on “academy controlled infrastructure” and some of the questions that our community engages in pursuit of that control.  As we continue our strategic visioning work, and develop plans for future upgrades to HathiTrust services and systems, we’ll be considering many of those same questions so that we remain member-aligned and values-driven as we improve resilience in our services for the community.