Network Performance Management

WildPackets offers industry leading network performance management solutions.

An accurate picture of key network metrics as they change over time can be an invaluable tool. Not only is this essential for resource planning, but it can also help managers to head off trouble before it starts in earnest. This can be as simple as identifying new applications that may soon be using enough bandwidth to have a serious impact on network performance. Fluctuations in the network resources used by a particular process may be transient, or they may signal a real change in user behavior. Good trending statistics can help to distinguish one case from the other, and help network managers plan accurately to meet user needs.

Network Performance Management through Monitoring and Problem Detection

A large part of the network manager’s job consists of watching for variances from normal. Not every anomaly is a problem, but every problem stems from an anomaly (or at least, that’s the ideal).

WatchPoint, in conjunction with OmniPeek Network Analyzers and TimeLine network recorders, make it easy to manage your network performance, day in, day out. Network statistics such as over-all utilization, counts of nodes and protocols, error rates, and so forth provide a quick overview of basic network functions. In many environments, managers may want items such as these to be displayed continuously or refreshed at frequent intervals throughout the day.

Individual applications or processes may also require a continuous or very frequent update and display, particularly at times and places where these processes are heavily loaded. DHCP and email servers, for example, are often under their heaviest load for a brief period at the start of a work day. Omnipliance network recorders capture statistics on particular processes, as well as on the network as a whole. And OmniPeek Enterprise automates the process of collecting, displaying and saving that information. For example, you could use OmniPeek to automate the capture and calculation of statistics on all DHCP traffic on a segment, then save the statistics as HTML files every five minutes between 7:45 a.m. and 8:30 a.m., allowing you to check the most current DHCP statistics from any connected machine.

The Importance of Detecting Network Problems

The object of continuous or frequent network monitoring is to spot problems as they occur. Sometimes a user complaint about lost or degraded service is the first sign of trouble. This is not what network managers want. Ideally, they would like to be able to identify and correct problems before users notice through effective network performance management. When a problem is identified by a user, it is still necessary to refine the definition of the problem, and a network analyzer like OmniPeek are crucial tools in that process. There are a great variety of problems that can come up in any LAN or WLAN.

One example of a acute problem is the complete failure of a server or application process, or degradation to the point that workflows are seriously impaired or interrupted. Unlike intermittent problems, they don’t (as if by magic) appear to resolve themselves. Unlike chronic problems, they can be resolved immediately.

Alarms are an excellent tool for watching for signs of failure or serious stress in key network and application processes. OmniPeek lets you set two-stage alarms on any statistics item. Statistics in OmniPeek are calculated on a full range of protocols covering all the leading network protocols and many of the most widely used applications, including SQL (multiple versions), and VoIP, video and wireless.

Find out more about VoIP Monitoring and Analysis.
Find out more about Video Monitoring and Analysis.
Find out more about Wireless Network Analysis.

In OmniPeek the Expert function can diagnose particular types of failure in a variety of important application and network processes. The Expert not only identifies problems as they occur, but makes it easy to immediately identify the source of trouble. The Expert can even suggest typical solutions to the problems it diagnoses.