The SLA angle
|
|||
 | |||
|
Service-level agreements exact financial penalties when networks go down, but they don't indemnify customers for lost business. So they can be pretty meaningless in the event of a major network outage.
"If our network is not up 99 percent of the time, we get the whole month free," says Al Perry, vice president of information services for Texas Industries, a $1.5 billion producer of cement, aggregates, concrete and steel products based in Dallas. The company needed a new WAN to connect more than 50 locations spread across the country, and outsourced both the implementation and ongoing management to NetSolve.
"Such guarantees wouldn't help much if your network were to go down for an extended period of time. You don't want your money back; you want your network to stay up."
However, the details of the SLA can tell you a lot about service providers and the performance levels to which they feel they can commit. SLAs shouldn't be covered in a cloak of legal caveats, and they need to focus on how available your network is, not how quickly a problem gets reported.
Make sure "response" doesn't just mean opening a trouble ticket, and be careful to distinguish between "restore" and "resolve." Restoring is a quick fix to get the network back up, while resolving involves changing things so that the network doesn't go down again.
Reporting is an important part of SLAs-what type of reports you will get on the performance of your network, how often you will get them and in what format, and how readable they will be. Some service providers have only standard SLAs, while others will customize them to some extent.
"You might want to subscribe to different response times at different locations," says Perry. "For example, our Virginia steel plant has much higher uptime requirements than our California cement plant."
Send this article to a colleague
