>> Dan: From the standpoint of your infrastructure, obviously, you have to deal with a few days where you have some extraordinary peak usage and how are you managing that?
>> Well, that's a huge challenge. We need to build our infrastructure to manage the peaks, obviously. And then the rest of the year we have that excess capacity. So basically what we do is we scale our environment horizontally. When we need additional hardware of software, we scale the environment horizontally. We do that in three hosting centers around the country. And we build the environment to satisfy the peaks. When we don't need the environment for the peaks, we expand our development environment and we start virtualizing, so we're tapping into those resources more and more so that we use that excess capacity to prepare for the peaks and we make more efficient use of that technology as best we possibly can.
>> Dan: So you're not outsourcing in terms of using an external service when you need peak hours, as you might for an electrical utility, you need more power, you simply go to the grid and get more.
>> Right. We're not doing that now. We're, of course, exploring that in a variety of ways. But the applications that we're supporting are so mission critical to us during our peaks, that we're a little hesitant to do that. We're absolutely pursuing that as well as a variety of options to see what we can do to help better manage our peaks. Not the least of which is to even out our revenue stream over the year.
>> Dan: So what kind of utilization are you getting on your servers and storage given that you're built for peaks but those peaks are fairly rare.
>> Well, we could peak at 70 percent, 80 percent. We monitor our peaks very, very carefully obviously. We want some head room during our holiday periods. So even off peak we're monitoring everything. Every single component in our environment is being monitored and analyzed on an ongoing basis, performance, response time with respect to performance, capacity on each of our servers, utilization of all of our servers, so we do monitor that very, very carefully Dan, but we're pretty careful not to exceed about a 70 percent peak, I would say.
==== Transcribed by Automatic Sync Technologies ====
