With more and more enterprises both small and large rolling out Voice over IP (VoIP) solutions it has become necessary to do more specialized fault monitoring and network monitoring. Companies are rolling out VoIP as they are looking to cut costs and use a single network to carry both their voice and data traffic. When rolling out VoIP, companies need to ensure they have enough bandwidth on both the LAN and the WAN network, using a fault management tool combined with a mib browser can assist in this process. Without the proper allocation of bandwidth rolling out a VoIP service will lead to lots of failed or poor quality phone calls.

The use of fault monitoring and network monitoring becomes crucial to identify how much bandwidth is required to deploy the voice solution across the network and maintain adequate voice quality. VoIP deployments usually face issues with some sort of a network transport problem. Using a fault management tool to receive SNMP traps can assist in identifying network transport issues. Most companies are used to dealing with the deployment of a new data service where if a packet is slowed down by a few milliseconds the service will remain intact. However, in the world of voice a few milliseconds can mean hearing what the other person is saying and not. The four big areas where the difference comes in are delay, jitter, packet loss, and echo. One-way delay is said to need to be less than 250 milliseconds, when this threshold is crossed, the transmissions of voice overlap and it makes having a conversation difficult. Jitter measures the change in the difference between inter-packet arrival times and inter-packet departure times, the smaller the time difference the better the call will sound. While packet loss occurs when one or more packets do not reach their destination, in a voice conversation this is critical, so techniques have been developed to re-transmit the last packet. Some companies use a mib browser to check on the values of packet loss in the network, as this can help them identify issues. Finally, there is echo, which is the last test of the four items companies worry about, especially if there is no distinction between voice and echo. Proper fault monitoring and network monitoring will enable companies to correlate the values of each with the VoIP service to ensure a clear voice call. Also, one can set thresholds for these key metrics and have SNMP traps sent to a fault management solution for further analysis. This correlation also allows companies to fine tune their QoS policies and continue testing until the right amount of bandwidth is allocated to the VoIP service.

Monitoring QoS policies on an on-going basis is extremely important in a VoIP environment. These policies need to be checked to ensure the items discussed above are within acceptable ranges. Fault management system that receive SNMP traps can assist customers in ensuring metrics are within specified ranges. Usually when one or more of the factors are outside an acceptable range part of the QoS policy has been violated. Using a mib browser and SNMP trap receiver like ByteSphere Technologies OidView product can assist companies in rolling out VoIP services. OidView's mib browser has a built-in QoS profiler, which allows companies to browse and verify their QoS configurations and make changes accordingly. Also, as companies continue to roll out VoIP across the company there is inevitably going to be more and more calls, therefore requiring more bandwidth for the VoIP service.

On the WAN side companies are looking to maximize their underutilized links to carry VoIP traffic at no extra charge. With this said it is important for them to carry over their practices of fault monitoring and network monitoring to the WAN. If they do not have the ability to monitor the WAN it is important for them to have a Service-Level Agreement with their provider for WAN service they are purchasing or ask for a log from the providers fault management solution, showing the SNMP traps received for their links. Devices such as firewalls and others which do address translation can cause latency in the network, and for a voice call this could cause a call to sound choppy or have long pauses.

Fault monitoring and network monitoring is key to maximizing VoIP services. The idea is to use the smallest amount of bandwidth necessary for the VoIP service to run flawlessly. By using an iterative process and a mib browser, companies are able to minimize the amount of bandwidth needed for the voice service, while maximizing the use of their LAN and WAN links. This allows more bandwidth for other services to run across the same link, and therefore save the company money.