One thing often misunderstood about quality of service (QoS) guarantees is exactly where the guarantee starts and ends. Unbeknownst to end users, many such agreements are limited solely to the service provider’s premises, according to one industry observer.

Jenn Markey, vice-president of marketing for Kanata, Ont.-based Empowered Networks Inc., said the service provider tends to provide a higher level of QoS through the core, the backbone and the edge, but they don’t extend it out to a customer’s equipment. Generally, only on the largest of customer accounts do providers offer QoS to the last mile.

“And the way they’ll do that is (by) sort of putting together solutions, most often with some sort of probe device, to find out what the end-user’s experience actually is, and they’ll try to marry that into an end-to-end quality-of-service guarantee that would go way above and beyond what they normally do for their business market,” Markey said.

A run-of-mill QoS guarantee might offer 99.999 per cent or even 100 per cent uptime with many caveats around it, but it wouldn’t talk about the level of utilization in the network. QoS guarantees tend to be fault-based (is the network up or down?) rather than assurance-based guarantees (how well is the network working?).

So is QoS possible in the last mile? To achieve it, there has to be a significant investment on the service provider’s part on the access layer because access generally has a big traffic jam problem.

“If you want to do quality of service at the access layer, you’re talking about getting an environment like traffic shaping,” she said. “You have to throttle back bandwidth on the access, so you get into things like queuing and flow control.” She added that most providers don’t want to make that kind of investment.