Building an IT Framework to Improve Performance

Samuel Greengard Avatar

Updated on:

Managing any fast-growing business can present huge IT challenges. For Mavis Discount Tire, an independent chain with 160 stores and service centers in New York State, Connecticut, Pennsylvania and Massachusetts, managing everything from inventory and a call center to online appointment scheduling is critical.

“We are technology-oriented retailer,” explains CIO Edward Schwartz. “It is important that we meet increasingly demanding business requirements.”

The company, which has roots extending back to the late 1940s, has motored into the digital age. “We have a lot of database-driven applications and we have an enormous need for redundancy,” Schwartz says. “So, it was important to build an IT framework that supports our need for performance.”

Several years ago, the company turned to Microsoft Cluster Server. That way, “If a server fails, another will take over,” he adds. “Our Website and various services operate 24×7, so we need to be fault-tolerant.”

However, the clustering technology introduced another set of challenges, including how to handle storage. The conventional approach would have been to use a storage area network (SAN), but with two or more servers in a cluster that share data, “Achieving optimal performance requires attention to detail on some extremely expensive equipment,” Schwartz says.

What’s more, the approach introduces another point of failure. “Now you have a shared storage array that you need to make redundant,” he points out. “So you have added a substantial amount of equipment and additional complexity.”

Mavis Discount Tire began exploring ways to more efficiently manage IT within a clustered environment. It eventually turned to SIOS Technology’s SANless clustering software to achieve high availability for its business-critical MS SQL Server applications and databases.

The system relies on local storage instead of a SAN, and it provides real-time replication of data for better synchronization, as well as failover and business continuity. Since implementing the solution, the company has added physical and virtual servers and has achieved a high level of flexibility about how it uses them.

For example, “We can now say, ‘We want to move 10 servers and split them between location A and location B,’” Schwartz explains. “We can shut down one node of the cluster and move that cluster over to a LAN at another location for additional redundancy or for other reasons that might involve performance or where users are located.”

In addition, the IT department now has greater flexibility in terms of turning physical servers on and off. “In a mission-critical environment, we cannot afford to turn a machine off—or have a failure and wait 15 minutes for it to come back online,” he adds.

In the end, the SIOS clustering software has reduced the need for multiple SANs, dedicated and redundant external switches, and the associated IT administration and oversight that comes with a more complex IT infrastructure. In many cases, it’s now possible to roll out applications and servers, such as a VoIP phone system, without engaging in any additional design. The company simply purchases additional licenses, and the servers replicate the existing configurations and data.

“We make an effort to use information technology as efficiently as possible,” Schwartz says. “The ability to manage clusters and storage allows us to react to conditions and events quickly and expand when needed.”

 

Samuel Greengard Avatar