Archive for category Network Info

New DNS Clusters

We’ve recently increased our DNS server availability by creating three separate clusters of DNS servers around our network. DNS servers are queried every time you visit a web page, check your email or go online. They’re critical in making the Internet go.

At our primary datacenter facilities in St. George we have deployed two separate DNS servers in each facility for load-balancing, load-sharing and high-availability. This deployment method uses Cisco load balancing cards in our redundant Cisco 6500 Switches in order to deliver the high-availability and load balancing capabilities. In the event that one site is cut off from another, one facility can handle 100% of the DNS load. As load grows, we can easily add more servers at both data facilities to keep loads low and response times fast!

At each of our Cedar City and WebbHill (St. George) locations, we have deployed two DNS servers to help with load sharing at each location. Customers in Cedar City will automatically query our local Cedar DNS Servers, while those on our Licensed Wireless in St. George will query the WebbHill servers. The great thing about this setup is that if one server fails the other will take over for the load. If both are disabled or not responding to DNS queries, our main cluster in St. George will handle the load automatically.

What this means to you is increased availability in DNS resolutions. This also allows us to do maintenance on one portion of the cluster without causing wide spread problems.

Next up will be our deployment of a front-end mail server cluster.

More on that later.

, ,

1 Comment

Scheduled Maintenance – 10/10/09

cisco_6509InfoWest is scheduling a maintenance window on October 10th at 3am. We plan on upgrading our Cisco 6509 in our Tonaquint Datacenter POP next Saturday (10/10/09) at 3am. We’ll be installing new redundant supervisor modules and re-grooming fiber.

This will affect one of our redundant paths to the Internet, but all InfoWest traffic will automatically re-route around our metro fiber-optic gigabit ring and go out our second path.

Plan on up to 1-hr of downtime for fiber connections and co-located servers at Tonaquint during this window, although our upgrades should not take longer than 15 minutes before fiber and co-location circuits become reachable again.

, , , , ,

No Comments