It’s a fairly common scenario: users complain that the system is responding slowly; the network engineer checks port utilization stats, sees that during the busiest parts of the day the network port is at 50 percent utilization, and can’t understand what users are complaining about. In fact, there may actually be a network management problem that is being hidden by the granularity of the report.
The port is like a stadium turnstile; at any particular instant it is either being used or it isn’t, but you can’t get two people through at a time. Similarly, a port is either busy or idle, and like the turnstile, those who arrive when it is busy must wait their turns. A report that says that the turnstile was only 50 percent in-use during the hour before a concert, may conceal the fact that it went unused for the first half hour and served a long line of frustrated attendees during the second half.
In both cases—network port and turnstiles—you’d get a more useful measure by reducing your measurement intervals. That would highlight the intervals during which the resource is fully utilized and might lead you to install additional turnstiles or ports (or change the way users queue for the resource) at busy times.
If users and reports disagree as to whether there is a problem; there may well be a problem, and reporting intervals may be the first place to look for answers. Of course identifying a network management problem is only the first step, and while adding resources is one way to solve the problem, there may be other approaches. Queueing theory addresses various ways in which resources can be deployed to serve users. As network complexity increases, so does the number of options available for balancing costs and service levels. And while the simplest options may not be the best, neither may the most expensive. Consolidated Communications offers consulting and data management services that can help you balance service levels and cost and possibly improve both.