The trigger might have been outwith their control but everything that happened afterwards was a direct result of AAISP mistakes.

The BlueSquare report clearly states that dual-supply racks were unaffected. This is really basic stuff - you double up absolutely everything. For whatever reason AAISP had equipment which only had a single supply.

Next, although the main routers are a resilient pair, they both seem to have been configured to use the same method to determine their availablity (i.e. connectivity to Maidenhead). So, another single point of failure wiped out both routers.

Next, the router configuration was clearly wrong and had never been properly tested in that this quagga router took over when it was not intended to. It is ironic that the router that stayed up is to be replaced (presumably by one of the ones that failed). However, even tho this router stayed up, it did not seem to be able to route to very much. I certainly could not get outside the AAISP network even to UK sites.

Lastly, it seems that the office was unable to properly monitor and diagnose the fault quickly because they were dependant on Maidenhead somehow and their contingency (GPRS) was not up to the job.