Performance Management

Performance management is used for trending and capacity planning. Data is collected and analyzed . Network engineers and managers review it, looking for trends that may indicate the need to increase or decrease network capacity. Link utilization, Frame Relay FECN and BECNs, and router CPU are some items that may indicate a change in capacity is required. Response time, measured regularly between routers on the edges of networks, as close to end users as possible, directly shows the impact of the over- or underutilized network components . It also shows the improvement when the capacity has been modified.

The performance management station is continuously collecting data. It collects data via SNMP, polling for groups of variables on regular short intervals, such as every 5 minutes. The data is stored in raw form for historical research. It also is processed and reported as minimum, maximum, and average values on an hourly basis. The data may be processed during the day for previous hours, providing almost up-to-the-minute reports and graphs, or it may be processed at some time during the night, providing reports detailing the preceding day's numbers .

The system also should report on the amount of time during the day that the values fell into certain ranges.

Consider link utilization, for example. It is helpful to know how much time during the day the utilization fell into the following ranges: 0 “20%, 20 “40%, 40 “60%, 60 “80%, 80 “90%, and 90 “100%.

After a few days of collecting and processing data, the system processes the data further, reporting on minimum, maximum, average per day, and ranges for the time period. If the time period is a week, the minimum, maximum, and average should be reported for each day, and the amount of time during the week spent in each range should be reported.

The performance management system needs to be up and collecting data uninterrupted. Trend analysis is meaningful on the collected data only after collecting data for an extended period of time. Projecting the trends into the future helps to determine when breaking points are likely to occur.

Flexibility makes the performance management system more valuable. Configurable time periods in which to view data, with reports as near to real time as possible, make this a very valuable tool for quickly identifying the need for capacity changes.