Network monitoring is the information collection function of network management. Network monitoring applications are created to collect data for network management applications. The purpose of network monitoring is the collecting of useful information from various parts of the network so that the network can be managed and controlled using the collected information. Most of the network devices are located in remote locations. These devices do not usually have directly connected terminals so that network management application cannot monitor their statuses easily. Thus, network monitoring techniques are developed to allow network management applications to check the states of their network devices. As more and more network devices are used to build bigger networks, network monitoring techniques are expanded to monitoring networks as a whole.
As more people communicate using networks, networks have become bigger and more complex. The proliferation of the internet has increased the pace of network expansions. At this age of big and complex networks, network monitoring applications need to use effective ways of checking the status of their networks so that network management applications can fully control their network and provide economical, and high-quality networking services to the users. It is very important to know what are the goals to achieve in network monitoring. By knowing the goals of network monitoring, network monitoring application can choose among network monitoring techniques that will best help them monitor their networks.
There are generally three basic goals for network monitoring;
- Performance monitoring
- Fault monitoring
- Account monitoring
-
Performance monitoring deals with measuring the performance of the network. There are three important issues in performance monitoring. First, performance monitoring information is usually used to plan future network expansion and locate current network usage problems. Second, the time frame of performance monitoring must be long enough to establish a network behaviour model. Third, choosing what to measure is important. There are too many measureable things in a network. But the list of items to be measured should be meaningful and cost effective. This list of items to be measured is called network indicators because they indicate attributes of the network. Here is an example list of network indicators in Table 1.
Performance monitoring deals with measuring the performance of the network. There are three important issues in performance monitoring. First, performance monitoring information is usually used to plan future network expansion and locate current network usage problems. Second, the time frame of performance monitoring must be long enough to establish a network behaviour model. Third, choosing what to measure is important. There are too many measureable things in a network. But the list of items to be measured should be meaningful and cost effective. This list of items to be measured is called network indicators because they indicate attributes of the network. Here is an example list of network indicators in Table 1.
Table 1: A list of network indicators.
Network indicators
Description
Circuit Availability
The actual time that a user can dial up to a network and the network connection is available for the user
Node Availability
The actual time that a user can use network nodes, multiplexers and routers without having error.
Blocking Factor
The number of user who cannot access the network because of busy signal in theory.
Response Time
The time to transmit a signal and receive a response for the signal.
Fault monitoring deals with measuring the problems in the network. There are two important issues in fault monitoring. First, fault monitoring deals with various layers of the network. When a problem occurs, it can be at different layers of the network. Thus it is important to know which layer is having problem. Second, fault monitoring requires establishing a normal characteristics of the network in an extended period of time. There are always errors in the network but when there are errors, it does not mean the network is having persistent problems. Some of these errors are expected to occur. For example, noise in a network link can cause transmission errors. The network only has problem when the number of errors has suddenly increased above its normal behaviour. Thus, a record of normal behaviour is important.
Account monitoring deals with how users use the network. The network keeps a record of what devices of the network are used by users and how often they are used. This type of information is used for billing user for network usage, and for predicting future network usage.