One of the traditional roles of a network operations center has been to provide fault management primarily for wired networks. Part of this traditional role includes functioning in an environment in which the end point devices are either a desktop computer or a laptop.
The role of the NOC is changing in part because the IT environment is changing. For example, while desktop computers and laptops still function as end point devices, in many instances so too do VoIP phones, softphones, PDAs, and smart phones. In addition, these new end point devices are accessing a wide variety of new applications that may or may not have been sanctioned by the IT organization. A more important aspect of the changing role of the network operations group is that while these groups are still responsible for providing fault management, the majority have also assumed responsibility for managing application performance over both a wired and a wireless infrastructure.
As NOCs are finding out, application management is notably more difficult than fault management. One of the factors that make fault management relatively easy is that a fault in a switch or router often leads to an outage, which is readily noticeable. Another factor is that it is easy to set alarms to indicate that some component of the IT infrastructure has failed. That being said, one of the primary challenges associated with traditional fault management is that a given outage will usually generate numerous alarms. Typically one of the alarms is associated with the actual outage.
For example, if an interface on a router is defective, it will generate an alarm. However, alarms will also be generated because the WAN links that terminate on that router interface are not functioning. In this example, the alarms on the WAN links were not generated because there were no problems with the WAN links. Rather, these alarms are symptoms of the fundamental problem being that the interface on the router is defective. Hence one of the challenges associated with traditional fault management is suppressing all of the false alarms that are associated with the symptoms of the problem and not the problem itself.
We should point out that we are not saying that fault management is easy. As one IT organization recently explained to us, “We still spend a lot of valuable time chasing down IOS bugs.” What we are saying is that as difficult as traditional fault management is, application management is notably tougher. The next couple of newsletters will discuss this further. In the mean time, more information can be found here.