Summary

VoIP networks leverage multiple protocols, and this chapter introduced such protocols as H.323, MGCP, SIP, RTP, and RTCP. VoIP components can be deployed using one of several call-processing models, such as a centralized call-processing model, where a cluster of Cisco CallManager servers reside at a central location with remote IP phones registering with the CallManager cluster over the IP WAN. Conversely, a distributed call-processing model might have a Cisco CallManager cluster at each geographic location.

A VoIP network's scalability depends on a scalable dial plan design. This chapter examined various dial plan considerations and discussed how the North American Numbering Plan (NANP) might be adapted to a corporate environment.

When voice traffic is added to an existing IP WAN link, previously only being used for data, a designer must provision sufficient bandwidth to support anticipated call volumes. This chapter delved into the mathematics of calculating the amount of bandwidth required to support a voice call.

This chapter also covered the fundamentals of traffic engineering, the science of calculating the number of trunks required to support a certain percentage of calls during the busiest hour of the day for a phone system. In the VoIP world, traffic engineering goes a step further by equating a number of trunks with a bandwidth amount.

Finally, this chapter explored security issues surrounding VoIP networks. For example, existing firewalls might inadvertently block ports required for call setup. Also, while VPN networks enhance the security of VoIP sessions, VPN technologies add overhead to each packet. This chapter described how to account for this extra overhead and how to allow appropriate traffic through a firewall.