As IT departments and the communications industry move to a single converged voice and data network solution, the demand on the network to provide specific service levels increases. This demand for higher and differentiated service levels for certain applications has created a need for a reliable and scalable Quality of Service (QoS) model in the network. Prior to the introduction of Voice over IP (VoIP) into a production network, the environment must be equipped to support the reliable transport of voice and a solid end-to-end voice quality plan must be defined and in place.

Typical demands on the network when supporting voice are continuous uptime (99.999%) and consistently low deviations in latency (jitter). As a real-time application, VoIP is far more sensitive to latency, jitter, and packet loss than standard data applications as users have no tolerance for garbled or broken speech. VoIP continues to be compared to traditional telephony in terms of voice quality and reliability as this is the level of service quality to which most users are accustomed.

QoS is a descriptive for defining how IP packets are dealt with through network devices. Mean Opinion Score (MOS) is used to measure the "quality" of a telephone call. QoS on the network helps facilitate a better MOS score for voice on an IP network. QoS is not a standard or protocol, but simply a generic industry term for outlining technologies, standards, and strategies to provide for network quality. In general QoS, to facilitate good voice quality and high MOS, requires that packets carrying real-time voice traffic cannot be delayed and must be prioritized over data traffic, which can better tolerate being slightly delayed. Most jitter in the network is caused by queuing delays associated with momentary or chronic congestion. QoS for voice can help make this queuing delay transparent to the voice packets.

This paper examines a case study based on Intel's experience of deploying VoIP [2] with voice quality within a campus and converging voice and data on the existing local area network (LAN). The methodology used to enhance end-to-end voice quality includes ensuring bandwidth, enabling QoS and optimizing IP telephony endpoints. This case study demonstrates that even a basic voice quality and QoS strategy can produce high-quality results with minimal infrastructure upgrades.