It's been a while since we've seen a reader question, but they've begun trickling in again.

In doing some due diligence on figuring out some possible VoIP implementations for our corporate set up, just had a general question for you -- what are the key things to look for in managing a VoIP network? how do i manage this on a daily basis?  your thoughts are highly appreciated.

In order to effectively answer this question, we really need to explore more deeply. One fo the key aspects of managing an enterprise VoIP network is based on the thought process that went into the early planning and design. We need to understand what tools and resources were implemented it the development of the solution. These help us understand what we're going to manage.

First and foremost there is the basic network management component. This is often provided by your enterprise network operations center and focuses on the most basic health of the network. Monitoring of network elements often focuses on two key areas - uptime and connectivity. These are frequently monitored using Simple Network Management Protocol (SNMP) and basic ICMP tools like ping.

Tools like Vital Suite, OpenView, SolarWinds Orion, and such monitor network elements to ensure they're up and running. They also provide ping testing to ensure the network connectivity between nodes is up and providing acceptable response times. These tools vary in complexity and capabilities. They certainly offer much more than these basics. They can monitor percentage of utilization on links, provide alerting to status changes, and a wide array of other health-check information.

When deploying VoIP, there are three parameters that are absolutley crucial for monitoring -

• Delay or Latency
• Packet Loss
• Jitter

Unlike protocols like HTTP, VoIP protocols typically use User Datagram Protocol (UDP) for transport. UDP is a connectionless protocol designed for quick delivery. There's no mechanism for retransmission of packets that were lost or degraded. This conscious decision was driven by VoIP's real-time nature. The process of retransmission using TCP is quite overhead intensive. The TCP three-way handshake induces delay that actually impedes the delivery of high quality voice.

VoIP requires very low delay because it's a real-time, two-way communications stream. If the network latency is too high, voice quality quickly degrades, Voices sound artificial, or choppy.

Packet loss can be tolerated to a small degree, but high pack loss results in parts of words, syllables, perhaps even entire words or phrases being lost in the conversation. Extreme packet loss results in degradation of voice quality to the point it becomes unusable.

Jitter has to do with consistency of packet delivery in the network. Network elements can accommodate some degree of jitter through buffering, but excessive jitter results in bursty delivery of packets and a choppy quality of voice delivery that users just can't tolerate.

There are two VoIP network management solutions that I'd consider to be the premier, or flagship services in managing VoIP networks - Prognosis from Integrated Research and NetIQ's Systems Management suite, including Vivinet Diagnostics. These two producte represent the best tools in VoIP management to effectively manage, not the IP network elements, but the VoIP service and devliery aspects of those elements. These tools provide the view of your network that allows you to understand what's going on that could negatively impact call quality and service delivery.

You ask how to manage on a daily basis. You must be proactive. VoIP service rides on top of the IP network. A denial of service attack against a web server causes network congestion. Congestion can easily increase packet loss. Any number of fairly common IP problems, that may not bring the data network down, can degrade VoIP service to the point of unusability. The only way to successfully deliver VoIP is through vigilant, proactive, ongoing monitoring.

Technorati Tags: VoIP, VoIP Management