At around 11.30 PM last night the University's core switch failed. This will have resulted in a campus-wide network outage and will have affected both Internet and local area network connectivity. Connectivity was restored at about 7.20 AM this morning.
This failure was almost identical to the one of 2005/08/13. A similar failure occured at about lunch time on Saturday but was short lived (it lasted under ten minutes). Whilst the direct cause (trigger) of these failures is not known, the symptoms resemble those of a known denial of service exploit against the switches. At present we're still awaiting delivery of a hardware upgrade that was ordered after the first failure. This upgrade is required in order to allow us to upgrade the firmware on the switches and thus address the issue.