Background
In recent years ITS has invested heavily in equipment and facilities to ensure continuity during
disruptions due to severe weather and other events. An industrial‐scale power generator, fueled by
natural gas, was installed at Head Hall in 2012, and has been kept in a state of readiness ever since.
When in operation, the generator keeps large batteries charged, which in turn power the networks,
servers, applications, and communications infrastructure housed in the main data center. Last year,
UNB’s IT disaster recovery facility was commissioned at the Wu Centre; it is on a different power grid
from Head Hall, and ensures the university’s IT needs can continue to be met in the event of loss of the
main data center. Furthermore, each of the 80+ buildings on campus has at least one IT networking and
data room or closet; every single one has back‐up batteries that either run until exhausted or are
connected to building generators, in which case they can run indefinitely.
Flawless performance
Post‐tropical storm Arthur barreled down on Fredericton unexpectedly and with a punch that left the
city and surrounding areas reeling (post‐tropical means that a storm is no longer strong enough to be
considered a hurricane, but by no means does that imply it is not dangerous!). Widespread power
outages lasted up to 9 or 10 days for many. The Fredericton campus was up and running normally within
a couple of days‐‐it could have been a lot worse. One of the key successes contributing to the rapid
recovery was the performance of UNB’s IT services and systems, which were completely operational
throughout the storm and the aftermath.
When power to the campus went out around 8:00 am Saturday, July 5, the Head Hall generator started
automatically, feeding power to the batteries that power the data center. The switchover was
instantaneous; no loss of service or data occurred. Though power to the Wu Centre also went out, the
disaster recovery facility there also has an array of back‐up batteries to ensure continuity, and they took
over the power load immediately. Unlike Head Hall, the Wu facility does not have a generator to keep
the batteries going‐‐in this event it didn’t matter though, since power was restored to the Wu section of
campus by noon.
Lessons learned
Continuity of IT services during an event like Arthur is essential. For example, websites, email, and data
processing capabilities were unimpaired, meaning it was straightforward for university staff to post
updates and communicate effectively. VOIP phone services performed as designed, and in fact exceeded
expectations, but as back‐up batteries in some buildings depleted, phone service was lost. Of course the
university was closed at the time, but the limits of batteries became abundantly clear. In addition, when
power was fully restored to campus on Monday, almost all building networks and VOIP phones restarted
automatically, but a few didn’t, posing questions for ITS to research and resolve. There were also some
unforeseeable events—for example a generator at one building failed to function properly, but didn’t
impact the network and VOIP phones there. Network components that didn’t restart automatically were
manually restarted, and by noon on Tuesday IT services on campus were 100% operational, a
remarkable outcome after a serious event whose effects persist throughout the city and the region.
July 16, 2014, Terry Nikkel, AVP, ITS