 Quality of service has been adopted by ITUT as a framework for next generation networks. The classical understanding of QS provisioning is at different levels. We call them the delivery mechanisms, we look at those, then we will see how ITUT defines the framework, specifically it is a criteria that we shall totally see. So the QS delivery mechanisms in IP based networks are actually provided at different layers. For instance, if you look at the layer 2, the data link layer, these mechanisms actually allow the management of admission control and signaling at different access technologies in terms of the resources which are available at layer 2. For instance, the send and receive buffers which are available at the second layer. These layer 2 mechanisms are actually limited to the LAN segment or the network segment. So we can think about layer 2 technologies which change end-to-end. For instance, we have mobile network, we can have a wireless network and we can have an Ethernet all connected in tandem in cascade and an end-to-end IP connection is established as a service flow. In that case, we would be expecting that three different kinds of layer 2 QSs shall have to be provided. Now, in order to have consistent experience at the above layers, layer 2 QS for all these different technologies have to be managed in such a way that a user gets a seamless experience. Then at the network layer, we have end-to-end QS provisioning. It is known as the IP layer QS provisioning. Well-known architectures from the IETF are the differentiated services architecture, the integrated services architecture and the legacy multi-protocol label switching. Now, these network layer QS provisioning mechanisms have their own pros and cons. It is beyond the scope but these architectures are worthwhile to be studied and considered. So when ITUT adopted a QS framework, it did it as a standardization activity because there are other organizations at the regional level, even at the global level, which are providing QS to their networks but for NGN, some kind of universally accepted QS framework had to be designed. Now, ITUT is providing the lead but it is also mindful of the protocol evolution which is taking place in the ITF. So whenever some protocol modification or release takes place by the ITF community, so it is incorporated as a mandatory requirement by ITUT. So the QS criteria and parameters are defined by ITUT in a document which is known as the recommendations document g.1000. It is a general QS framework not very specific so it has certain criteria. There are seven criteria which are to begin with very general but provide a good holistic view of in fact complete view of what QS has to look like. For instance, it defines speed of different service functions, the response which is expected out of each functional element, then the accuracy in terms of if it is an audio call, the speech quality, success, a probability of a call, the billing which takes place for that call duration. In terms of availability, for instance, the coverage of a certain service provider in terms of geographical area and in terms of service providers, then service availability as such is also a QS concern because a service may be available very effectively in one service provider's jurisdiction but it is not available in another so the QS criteria says it is not a very good service. Then we have reliability. Reliability is basically the trustworthiness of the network in terms of how many calls it drops, how many calls it makes and the overall stringent billing requirements. I mean if the customers are complaining of over billing then it means it's not a very reliable service provider. Then in terms of security, obviously the privacy is a concern, fraud prevention of any kind has to be ensured. Then simplicity is basically the ease of doing business that is if a network has to undergo a software upgrade then it should not cause service disruption and network unavailability. Then if a certain service provider is asked to terminate a service then it should not cause unnecessary delay and procrastinate. It should actually be a simple procedure through which the network can be engaged and network can be disengaged for a certain service. Then we have the flexibility. Flexibility actually implies the ability of the network to change so that the user does not feel it has not been provided the grade or the quality of service as requested. So whenever there's a change in contract, if there's a clause that is introduced at one point in time, for instance if a user wants the prepaid service to be converted to a postpaid service then the network provider has to incorporate this into the user billing automatically and obviously the user should be provided the flexibility to pay through any means, for instance online versus offline. Not only the criteria are being defined but the service functions are also defined. The service functions should actually provide some kind of variation. If there is no variation then we can't ensure that quality of service is provided by a certain network. So in terms of service management for instance these services or the products or the applications which are offered have to have certain variety. We call it sales. Then the pre-contract user experience should be provided to make sure that a user has been exposed to the kind of service it is going to get. Then the provisioning of service has to be provided in a variety of ways. Some kind of modification or alteration has to be incorporated. Then throughout the provisioning of service and the contractual period for which a user is getting a service, continuous technical and customer support has to be ensured. If there is a disruption of service or some kind of inconvenience it has to be dealt with through proper repair regime. And lastly when the user wants the service to be terminated it should be given different opt-out models. Then there is also a flexibility expectation so that quality of service can be provided in terms of the information transfer. Information transfer means what all information and what levels of information can be provided in terms of connection. For instance connection establishment connection termination should be logged somewhere. Now this level of logging has to be different so that different kinds of information can be retrieved and business logic can be developed depending upon what connection information is available. Then billing and the overall configuration control of the user should also be flexible so that if a user experiences certain billing format and user is interested in getting billing in another format it should be provided. And if a user has a certain graphical user interface priority then it should be allowed to tailor or alter the graphical user interface correspondingly.