The communication requirements for a worldwide spectacle like the Winter Olympics can mean a network undertaking as massive as organizing the games themselves. With the added task of building out the Olympics' first all-IP network, Avaya spent more than two years designing and testing a network that would serve about 35,000 people. Dean Frohwerk, Avaya's chief architect of the network, discusses what went into creating the Olympics network.
CTOEdge: What were the requirements for the Vancouver Olympics network?
Frohwerk: The original requirements were about resiliency and reliability-the Vancouver Organizing Committee required 100 percent uptime, with no outages or downtime. We worked with Bell Canada and the Vancouver Organizing Committee for the 2010 Olympic and Paralympic Winter Games (VANOC) to enable that network design and we think we've put one together that is one of the most tested around the globe. It literally has been tested for years -- we got started January 2007 and since day one we've been working on the network architecture, looking at the requirements, determining the equipment needs, etc. In September 2007, we got our test labs up and running, and we've been testing ever since.
'We tried to adopt a one-team philosophy when providing services, which means we've been able to provide greater bandwidth and more efficient use.'
CTOEdge: What was the biggest hurdle in provisioning such a network?
Frohwerk: This is the first all-IP Olympics on one converged infrastructure-we are delivering VoIP over an all-Ethernet, all-IP network. We needed one converged network and offer all services on that network. That took a lot of testing.
The architecture was one of the areas where we get the most lead time, and it has proven to be sound and we've been able to implement and expand out to all the different venues. There are close to 30 venues and 70 support locations.
One major challenge we encountered was late access to some of the venues-the Vancouver Canucks just left [their arena stay] three weeks ago and so we had a short time to turn that arena into an Olympic venue. We couldn't shut down the existing network for the Olympics, so we ended up running two parallel networks in there. We had to do a full network build in a short time.
But the overriding hurdle, I would say, was the scale-we spent a lot of time building out models in the labs, and we did full stress- testing. We need to be able to handle 35,000 users starting today, so we spent a lot of time testing to ensure we could.
CTOEdge: What is different technology-wise about this Olympics than Olympics past?
Frohwerk: The fact that it is such a large undertaking, and it is the first all-IP network is something. But it was also mandated that the operations center be converged as well. At the Beijing Olympics, there were a number of different communications vendors running their own networks. But we tried to adopt a one-team philosophy when providing services, which means we've been able to provide greater bandwidth and more efficient use. The venues in our model get 20Gigs of bandwidth, compared to 2Gigs at the Beijing Olympics. The order of magnitude in difference is enormous.