In the case of a VoIP or Video call it's not just one packet that has to be on time, every packet has to be on time, and in both directions.
The definition of time is not measured in seconds, but thousandths of a second!
So, what does this mean for network measurement...?
There are many significant metrics that must be measured on the network to assess that it is able to deliver good quality for audio and video.
The first significant VoIP metrics is Jitter.
Jitter defines the amount of time between the packets traveling in the same direction on the network.
Learn about JitterThe second significant measure is packets lost.
VoIP packets are never recovered. When a packet is dropped the call quality is threatened.
Learn about Packet LossThe third significant measure is the amount of bandwidth and how it is used.
Data packets are large and consume bandwidth.
Learn about VoIP BandwidthTo ensure quality MCS measures not only Jitter but also peak Jitter, discard Jitter, and average Jitter.
Equally important is packet loss. MCS measures not only packet loss but also the distribution of that loss.
MOS score is also reported, which is the popular metric for overall VoIP health.
Factors that have a big impact on VoIP Quality, such as available bandwidth, are supported by MCS's ability to customize the test for call lines and codecs, which consume bandwidth.
How many calls can your network support?
MCS provides a sophisticated VoIP Capacity assessment, which quickly validates the number of concurrent calls that can operate on any given connection while still maintaining audio quality. This test automatically stops at the bandwidth value where call quality metrics start to fail, namely Jitter and packer loss exceed the threshold defined for the test. This is a bit like putting cars on the road and stopping when cars fail to reach their destination on time. Such a test would essentially identify the exact traffic capacity that initiates the 'Rush Hour'.
The MCS solution allows you to easily deploy a VoIP/Video assessment portal for your own business. A powerful API allows for endless scope in terms of how the portal is presented to your end-users. Data can be easily segmented and the highly customizable reporting system makes analyzing results easily.
Testing your connection will uncover the true user experience.
Browser-based assessment testing allows you to present a customized portal to your end-users, which allows them to assess their home or work connection.
Satellite technology. Utilize our powerful MCS Satellites (software and hardware) to create testing points across your network. Continuous testing can then be set up between any and all end-points.
Jitter time will increase proportionate to the number of packets lost and audio content will be missing. The distribution of the packets lost is a critical measure.
Most VoIP tests report packet loss as an average, for example dropping 25 packets out of a thousand would be expressed as 0.25%. Sounds like a very small number and not worth worrying about.
However, consider the problem if the 25 packets were dropped contiguously. At 50 packets per second that represents 50% of the time window.
Jitter is measure in milliseconds and a good result is 0 (zero) milliseconds and a bad result is any number other than zero. The higher the number the bigger the threat to audio quality.
As a practical human example consider two cars that leave 5 minutes apart heading to the airport. If the cars arrive 5 minutes apart then the jitter value is zero i.e. no change in time. If the trailing cars arrive 6 minutes apart the reported jitter is 1 minute.
If the jitter time is large then the second car arrives too late and misses the flight. In other words, to deliver a good experience the car must arrive in time to catch the flight.
For example on 1Gbps workstation the smallest data payload sent from a browser is about 48 packets. An connection at 12Mbps will consume 1 millisecond of time for each packet. Why is this and why does it matter?
12 million bits per second is 12,000 bits per 1ms. A data packet is normally 1500 bytes which is 12,000 bits. Therefore 48 data packets will consume 48 milliseconds of time on the network. If your VoIP packets are behind the data stream then on a 12Mbps connect your jitter peaks will be at least 48ms or higher.
Remember 48 packets is the smallest payload of data, plus there are many other workstations that send data on the network. If jitter is consistently high then the jitter buffer will drain and cause the application audio quality to diminish. Most Jitter buffers are around 80 – 100ms and increasing the jitter buffer is not a practical solution because it increases the talk-over lag time which also affects call quality.