 Now, it is important to look at some quality of service aspects of various applications which run on top of the next generation networks. To begin with, Voice over IP seems the most eligible candidate for consideration because it represents in the most wide and generalizable manner the applications which run including multimedia, voice and video applications. So we'll start with developing the background for QS requirements, elicitation and then we'd look at a comparison between what are the voice over IP requirements in NGNs which are different from or similar to the requirements which have existed for a long time in public switched telephone networks based pure voice. So the journey starts from 1876 when Alexander Graham Bell in Bell Labs had the first experiment to have communication, wide communication for audio over telephone line. So that was the genesis and that resulted into the evolution of telephony. So we moved from public switched telephone networks, public land mobile networks which were basic networks requiring only the operator to worry about voice alone. So these were very original and patent voice networks. Then there was a requirement to add certain other services on top of those telephone voice networks. So that turned out to be more like a service. So it means voice was no more monolithically and by itself as a singular application related to that network, certain other services also emerged. So voice actually got a new look that is it was a service now running on a network. In NGNs the provisioning of voice is considered to be more like an application. So it means different applications can provide different voice related services. So it means that we have seen the migration from being a pure telephony oriented viewpoint on voice to what NGNs are all about IP centricity. This actually requires quality of service because in IP most of the traffic is best effort unless it is enabled to have priority and certain other US enabled features. There's a very interesting comparison between the network parameters and the performance indicators in PSTNs versus NGNs with regards to quality of service. For instance if you look at bandwidth PCM encoded 64 kilobits per second is an acceptable audio quality on PSTNs circuit space networks but in NGNs the bandwidth can be increased and more advanced coding schemes other than PCM can be incorporated to transmit very high quality audio. For instance in 5G networks for example the concept of crystal voice is used that is voice which is so clear and so high fidelity that it's crystal clear it's known as crystal voice. So bandwidth is not an issue because in NGNs any user who can subscribe to more bandwidth can have better audio but the problem is with the delay because of the statistical multiplexing in NGNs because there are no guarantees all the time so some kind of latency or delay is going to be experienced so delay as such and delay variation that is jitter are quite bothersome. So let's look at the parameter by parameter and KPI by KPI the differences between voice over IP that is provided on NGN and voice which is there on PSTNs. Let's start with bit rates so PSTN is locked on 264 kbps it is from the G.Rod 711 iDUT standard but as far as VoIP is concerned there is a great flexibility for example adaptive multi-rate is from 3GPP acquired from 3GPP can offer rate adjusting audio depending upon the network performance so voice over IP is way too flexible with regards to the bit rates. Similarly in circuit switched PSTNs since there is absolute circuit switching there is no flexibility to accommodate large number of users but here in NGNs we have statistical multiplexing so it results into accommodating more users with a certain degree of performance compromise but again that is so nominal or so meager that it can hardly be noticed at times. So this results into handling voice packets on NGN with the processing of their headers so including more headers like real-time protocol the family will study the user datagram protocol and the IP at the network layer now adding up these headers to the voice payload means that the original data rate that is specific to that particular encoding scheme increases so if we are talking about G.711 if you're using all these headers in packetized voice the effective data rate is going to be more than 64 Gbps per user then we have the delay in in in circuit switched PSTN networks the call establishment delay end-to-end including the switching delay including the signaling delay all included it does not normally exceed 150 milliseconds which is acceptable for human ear but when it comes to IP based NGNs there are a lot of processes which are going on with the packetized voice here like in multiplexing is taking place buffering may be required sometimes traffic shaping is required in certain networks which are resource staffed and have a lot of load and then certain priority and queuing and scheduling mechanisms are implemented all these incurred delays so it means that all these delays are going to add up to an extent where the audio quality is going to suffer in delay and delay variance so it means some kind of mechanism has to be introduced the quality of service we we've covered it in detail earlier can implement a high priority traffic scheduling so using the highest priority for audio it can always be put at the head of the line in a queue for being scheduled in at first this can result into delays that can go up to 400 milliseconds which is not acceptable but again this is the world of IP so that is that results into some kind of degraded performance but again using priority queuing you we can still bring it back to around 40 milliseconds then we have jitter again because of the unexpected behavior we can have variable delay for the arrival of audio packets on the receiving side so this can cause variable playback experience so you might have experienced this while talking to your loved ones you may have noticed that suddenly audio starts to get worse and it may even get dropped so in order to handle such a problem specifically to handle jitter one thing can be encoded in the network that is a qs-enabled path once established and can be used for all the subsequent packets which belong to the same stream of a certain audio conversation and then there could be the usage of a buffer at the receiver end in in the player such that if a player or if some audio system is playing back the audio on the speaker then it is buffered at the play out buffer known as jitter buffer for a while and then it is played back so it is buffered or enqueued then played this results into arrival of more packets which are then converted from from packet into analog voice to be sent on to the speakers this process takes time so by the time it is getting played back more packets would be received then we have a difference in the network equipment between PSTNs and VoIP enabled NGNs but in PSTNs there is a beautiful architecture of exchanges so having these exchanges results into well established and expected performance in VoIP this is carried out by something similar we have servers these servers run the session initiation protocol and diameter instead of signaling standard seven that's another obvious difference then convergence is something that means different different services all on IP actually result into sharing of resources because it's at the end of the day even if it is video or pure audio or it is a type or WhatsApp call it is essentially an IP packet this results into lower capital and operational expenditures this is not the case in PSTNs because PSTNs are all about audio so no comparison can be made here then PSTNs do not have a flexible or adaptive service architecture but NGNs do have that and lastly this availability or reliability of a system has to be an important consideration again PSTNs have five nines availability per year that results into or that translates into five minutes per year it's an excellent reliability figure but in in VoIP based NGNs they have low reliability because of certain features and one of them being that in PSTNs the exchanges and the telephone lines right up to your telephone set are powered by the operator but in voiceover IP the powering up of the end devices is the responsibility of the user so if there's no electricity in your home well you would not be able to get the voice quality or the voice service on NGN