Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
DQOS Exam Certification Guide - Cisco press.pdf
12.7 Mб

6 Chapter 1: QoS Overview

Foundation Topics

When I was a young lad in Barnesville, Georgia, I used to go to the bank with my dad. Each bank teller had his or her own line of people waiting to talk to the teller and transact their business. Invariably, we would always get behind someone who was really slow. (We called that Bubba’s law—you always get behind some large, disagreeable guy named “Bubba” in line.) So, someone who came to the bank after we did would get served before we would, because he or she didn’t get behind a “Bubba.” But, it was the rural South, so no one was in that much of a hurry, and no one really worried about it.

Later we moved to the big city of Snellville, just outside Atlanta. At the bank in Snellville, people were in a bigger hurry. So, there was one line and many tellers. As it turns out, and as queuing theory proves, the average time in the queue is decreased with one queue served by many tellers, rather than one queue for each teller. Therefore, if one slow person (Bubba) was talking to teller 1, when teller 2 became available, my dad and I could go next, rather than the person who showed up at the bank after we did. Figure 1-1 depicts the two competing queuing methods at a typical bank or fast-food chain—multiple queues, multiple servers versus single queue, multiple servers. The single queue/multiple servers method improves average wait time, but also eliminates the possibility of your good luck in choosing a fast line—the one with no Bubbas in it.

Figure 1-1 Comparing Multiple Server/Multiple Queue to Multiple Server/Single Queue



Multiple Q’s

Multiple Servers


Single Q

Multiple Servers

The bank in Snellville just chose a different queuing method, and that positively affected everyone, right? Well, the choice of using a single queue did have one negative effect—because there was only one queue, you could never show up, pick one of the many queues, and happen to get in the one with only fast people in it. In this scenario, on average everyone gets better service, but you miss out on the chance to get in and out of the bank really fast. In short, most customers’ experience is improved, and some customers’ experience is degraded.