For enterprise organizations, moving from a traditional phone system to IP-based telephony offers a number of concrete benefits. These include lower long distance charges and system maintenance costs, easier moves/adds/changes, improved integration with branch offices and teleworkers, and the ability to build distributed (virtual) call centers.

 

Deploying VoIP successfully, however, is a complex task that can be fraught with pitfalls. To avoid costly mistakes, network managers need to be sure that their IP infrastructure is ready to support the VoIP service, and also that the increased network traffic introduced by the new service will not negatively impact existing network applications. After deployment, continued monitoring and maintenance are crucial to ensure good call quality.

 

VoIP quality can suffer from various types of performance problems. These range from IP network problems (such as packet loss, delay, and jitter) to equipment issues (such as choice of codec, router and firewall configuration, and bandwidth allocation) to analog problems (such as echo and signal levels). Many of these problems are transient—occurring occasionally or only affecting certain calls—which can make them difficult to detect and diagnose.

 

Fortunately, most problems can be avoided with careful planning and the right performance management tools. To ensure successful deployment, network managers and architects should take the following basic steps:

 

 

Unlike traditional data applications, VoIP service is highly sensitive to IP transmission problems such as packet loss, delay, and jitter (variation in packet transit time). These can occur due to network congestion, equipment failure, and “route flapping.” In addition, the VoIP service itself creates increased network traffic, which can add to congestion and impact existing network applications. 

 

Many call quality problems occur in access links or on limited-bandwidth WAN or VPN links. If significant jitter or delay occurs on inter-site connections, this is a strong indicator that similar problems will occur during VoIP deployment. Budget bandwidth usage between sites and verify that routers can prioritize RTP traffic. For desktop connectivity, ensure that you use 100BaseT Ethernet architecture. Carefully examine Ethernet switch statistics for evidence of packet errors or excessive collisions, and upgrade equipment if necessary. 

 

 

 

Before deploying service, use a pre-deployment tool capable of generating simulated voice traffic with the traffic patterns and volumes expected from the “live” VoIP service. Because network conditions can vary throughout the day and week, testing should be conducted over a sufficient period of time to assess network readiness under a variety of typical conditions. Traffic load should be sufficient to simulate conditions similar to those expected post-deployment. (It makes little sense to test single VoIP calls if you expect twenty simultaneous calls.)

 

Testing tools such as Telchemy’s (News - Alert) DVQattest use installable software agents to both generate simulated test calls and perform real-time analysis of call quality and network performance. Such tools are well worth the expense, as they can locate problem areas before deployment and can also be used post-deployment for SLA monitoring and remote troubleshooting.

 

 

After a successful pre-deployment test has confirmed that there is sufficient bandwidth present, conduct a pilot trial of the VoIP service. Problems at this stage generally indicate an incorrectly configured router (or other network equipment) rather than a network performance problem.

 

 

As the deployment ramps up, carefully monitor any trends in key VoIP performance parameters, note any degradation in call quality, and investigate accordingly.

 

 

VoIP networks constantly change and evolve, due to changes in network and equipment configuration and variations in traffic. Continued monitoring and maintenance of the network is essential to maintain high quality VoIP service after the initial deployment.

 

The most accurate and cost-effective way to obtain real-time quality feedback on all user calls is using a distributed architecture of software probes—ideally those which support reporting protocols such as RTCP XR and SIP QoS, which provide a rich set of performance metrics (including MOS and R-factor scores) and can pass transparently through firewall routers. 

 

Lightweight, non-intrusive monitoring agents, such as Telchemy’s VQmon, can be embedded directly into VoIP endpoints such as IP phones and gateways, providing real-time call quality reports directly from each user desktop or handset.

 

The same monitoring technology is used in a range of mid-stream devices including probes (such as Telchemy’s SQprobe and Voyager), analyzers, routers, switches, and handheld test equipment. This gives network managers a range of complementary options for creating a scalable, end-to-end performance management framework for measuring call quality accurately and consistently.

 

Call quality reports from endpoints and mid-stream devices can be collected and analyzed by a central collection and mediation server. One example, Telchemy’s SQmediator, provides an interactive graphical view of current and historical call quality data, problem conditions and alerts, and integrates easily with existing NMS/OSS systems via SNMP.

 

Deploying IP telephony can be a daunting task—but by planning intelligently, testing thoroughly with the right diagnostic tools, and building a distributed framework for ongoing performance management, enterprise network managers can ensure a smooth transition and consistently high call quality after deployment.