With 20 parks and nearly US$1 billion in sales, Six Flags is the second-largest amusement park operator in the world. Since coming to Six Flags as part of a management reorganization two years ago, CIO Michael Israel has overseen a bottom-up rebuilding of the IT architecture in the parks and in the company's data center, which moved from New York to Dallas. Israel describes the amusement park business as a shopping mall with rides. "Spend per attendee is everything," he says.

How is the role of amusement park CIO the same as -- and different from -- any other CIO position?

The IT operations at Six Flags are a bit different in that there's a lot more operations involved here. We don't have a lot of that knowledge at the park level [because] we are a seasonal business. We are thin-staffed during the off-season, and there are times where the business is stretched in terms of people and operations knowledge. There are times when I'll be sitting in a park looking at how things are being done so I can get a better understanding of how we can reinvent that process.

What is your biggest challenge?

We're dealing with an outdoor environment. So something as simple as [wanting] another point-of-sale terminal in this location means that you're digging up the ground, laying conduit, laying cabling -- and your cost to get to that location is very high.

This environment is changing; it moves. New rides go up each year; retail stands get moved. Combine that with the fact that it is a seasonal business, and you have to take everything apart at the end of the year and put it back together on a very finite schedule.

How has the IT infrastructure changed since you joined Six Flags?

We've completely recabled our parks and laid new fiber. Everything is standardized. The data centers have been rebuilt with new cabling [and] new core switching, and the computer systems have been re-architected. About 70 per cent of the point-of-sale terminals have been migrated to HP, and we're working on the balance. We have about 3,200 POS systems and about 3,000 PCs in the network, and about 400 servers. We have Windows/SQL databases.

Is there any interaction between the IT systems and the ride systems?

No. Each ride is completely independent.

How did you split up IT operations between the parks and the data center?

Certain applications are at the park level, like our POS applications. Our core applications are in our data center, but necessary park applications are localized, because if we lose our wide-area network, the parks still need to be able to operate.

The park data centers are completely redundant internally, from our NetApp filers to our server farms. The farms are clustered, the servers use iSCSI boot, and there's dual power and dual network interface cards to everything.

How do you handle backups?

All servers boot up off NetApp boxes, and the data is snapped back to our central location, and all backups are done in our central data center. Before, we did local backups at each park, and those were stored at Iron Mountain. We eliminated the Iron Mountain contracts locally, plus I have a set of data at the park, I have a second disaster recovery set at the data center, and I have a third set that goes out on tape to Iron Mountain.