There are many elements to that question.
To be clear, technically speaking, the website was not the problem. The web service provider that's behind the website.... Our systems were functioning, but people could not access them.
There was a failure of a firewall that is set up by the private partner that provides the web services for us during the election. That firewall could not handle the spike we saw on election night—or throughout the election period, I should say. Throughout the election period, we saw double the visits to our website, and that was true as well on election night. They were more than double the number we had in the last two elections, when they were pretty stable. That caused the firewall to go into what we call “protection mode” and to stop the traffic from penetrating or to slow it down significantly, such that it was very difficult for most Canadians to get access.
There are a couple of things. One is that, at that point, we had to identify the source of the problem. We were very quickly informed, working with the Canadian centre for cybersecurity, that it was not a cyber-attack. This was almost instantaneous. We knew that this was not the case.
Our service provider, however, was not able to identify in real time the source of the problem. There are a few lessons learned from that event. Of course, from the service provider's point of view, they have augmented significantly the capacity of their firewalls. They've replaced the firewalls with much greater capacity, and doubled them.
We've also introduced protocols whereby we will be monitoring the pre-election tests more actively. Rather than letting the commercial partner do its testing, we will want to be more involved in seeing the tests that take place. The service provider is very committed to that.
We did have a couple of fail-safe solutions. One was that there's a separate channel that provides the results to the media consortium, so the results kept flowing through the media consortium to the media through a different channel. That's an important safety measure that we will, of course, maintain. Also, we have a replication of our website across the world on thousands of servers, except that there are two things about that replication. One is that it's a static website; we're looking into that. It provides information but not live information. In hindsight, we took too long to switch to that static website. We should have switched sooner.
We're drawing lessons from that event and making sure it does not happen again.