QoS denotes the capability of the network infrastructure, client applications and the end user terminals to deliver a service living up to certain quality levels. QoS requirements vary from service to service and depend directly on the specific services. In POTS, for example, there are detailed recommendations on QoS from ITU on maximum delay, blocking rate, MOS (Mean Opinion Score) etc.
QoS on the Internet is affected by a number of factors, including:
-
Delay
-
Bit Error & Packet loss
-
Speech compression
-
Echo
-
Firewalls
Different methods can be used to improve QoS. One can provide the necessary capacity in the backbone and access networks by ‘over provisioning’. QoS can also be implemented using one or more of following technologies:
-
Diffserve, ToS, RSVP, etc.
-
Using priority schemes in the IPv6
-
Using appropriate speech codes
-
Buffer size optimization
-
Packet size optimization
The main deployment of QoS is nevertheless connected to the introduction and development of IP version 6 (the advanced or next generation IP), which allows for end-to-end QoS provision.
In the managed IP infrastructures it is possible to provide measurable QoS, but this is more difficult in the best effort infrastructures like the Internet; however, in both cases regulatory measures may be necessary. An important issue is the facility-based operators' willingness to offer access to QoS provision to non-facility based operators. For example, a major debate in Europe and other regions is the lack of QoS provision in the wholesale Bit stream access products offered by the PSTN incumbents.