To transport language and video over data networks becomes more and more important for lot of different reasons and motivations. Unfortunately, the necessary circumstances as well as the potential dangers and inadequacy are less picked out as a central theme, so that it can come to considerable disturbances not seldom after the implementing of e.g. VoIP. It is often neglected how sensitive the "applications" language and video are – as a little tolerant they are compared with peak demand and high bandwidth workload and how important a quick time-critical throughput is.

While other applications also compensate more problematic circumstances thanks to TCP, the consequences are considerable with VoIP and video – and while one still forgives for the tough loading of a document, so impossibly is bad language quality, cracks and noise or even abnormal terminations of telephone calls with colleagues, superiors or even customers.

 

VoIP Readiness Check: When it makes sense and what it depends on?

One receives certified VoIP quality before the VoIP rollout – in spite of optimum configuration of the components – but only by an active testing of the infrastructure with synthetic VoIP conversations between different measuring points. Besides, the number of the concurrent talks is varied and measured at different day times and weekly times. The ascertained quality of the conversations reveals by the MOS value how many talks are possible at which times. A profound statement about the conversation quality can be expected before the real VoIP installation.

Hence a comprehensive VoIP Readiness Assessment (VoIP-premeasurement) should take up at least five working days. To sum up, it can be said that the VoIP-speech quality also depends on the behaviour of the users and the different real time applications (video Conferencing, client / server applications) on the network. Therefore, another active and passive monitoring of data and VoIP traffic is also obvious after the introduction of a VoIP system to guarantee proactively optimum performance and language quality.

 

 

Why is an end-to-end consideration of the Quality of service so important?

With the measuring / evaluation of the QoS (Quality of service) it is distinguished between passive and active measuring methods. With passive methods as for example the determination of a R-factor and the resulting MOS value (Mean Opinion score) according to the E-model (ITU T Rec. G.107) the so-called IP-network quality is measured in a measuring node. Network crossings and language codec conversions are not detected.

To be able to make a reasonable statement about the end-to-end language quality of a VoIP connection, active measuring methods must be applied as for example PESQ (Perceptual evaluation of speech Quality, ITU T Rec. P.862). With such procedures language reference files are sent. Then the language signals received on the other side are compared to the references. In this manner a statement about the complete distance incl. perhaps intervening networks (e.g., ISDN, SDH, ATM etc.) and codec conversions becomes possible what is not possible with a statement after the E model (MOS value).