 Now that we know what are the fine aspects and what are the differences between provisioning voice over IP on a packet space network like IP versus the PSTN, if you are interested in providing a very good quality audio on NGN then how QS can help in this regard and how QS is actually implemented by different network elements and different entities in the NGN functional architecture both at the transport stratum and the service stratum. So we look at quickly at the handover types and then we are going to look at how as a requirement a voice over IP can take benefit from the QS services available in the NGN architecture. So first and foremost once voice over IP is provided on fixed networks like which do not require any mobility the quality of service is essentially going to remain the same so that's not much of a concern. The problem actually arises when we have mobile devices because in mobile devices the attachment point or the access link actually keeps on delinking that is it keeps on changing and the attachment point actually moves from one network premises to the other so it means that the session has to continue in this case the quality of service consideration becomes very important. So once a mobile or a user equipment changes its location it's known as the mobility and mobility is handled through handovers if the technology from the departing network and the arriving network is the same it's horizontal and the devices which can the user equipments which can handle this are the single mode devices but if the departure and the arrival networks are different in technology then some kind of vertical handoff is required and then we have the multimodal or multimode supporting user equipment. Let's look at general architecture for QS enabled voice we have two operators since we are talking at a binary operation that is movement from one network to the other so we just have two operators both of these operators are showing different user equipments for instance we have LTE based user on accessing data we have a Wi-Fi based user and we have a mobile Wi-max user so you see it's generalized it is not very specific we have service session related control information that is service establishment service maintenance and service termination then we have whenever there's a change in network attachment then we have binding update and in order to run the binding update successfully and in order to use it some kind of discovery mechanism is required to discover the foreign network elements for that some kind of DNS query has to be run and then we also need some kind of network related control information that would inform the network elements that the new user is in proximity and has to be provided a service so all these are shown to you in different erode lines we see here the thickness actually is determining the difference between the two so just to recall from the NGN architectures perspective we have the network attachment control function which is closest to the network resources I mean closer to the physical layer then we have the resource admission control function that is the RSEF we have mobility management control function and service control function so you see here that the mobility management control functions are pairing with each other likewise the service control function are pairing with each other through DNS so it means that there is going to be a vertical relationship between the network elements and the functional modules in the engine architecture till the mobility management control function is activated and the mobility management control function is then going to coordinate with the peer mobility management control function in the foreign network and of course the service has to be provided through the service control function that eventually would also be talking to its own peer let's look at the establishment process in a step-by-step manner between different engine operators we have two operators operator one and operator two we have network we have user equipment one in operator one and user equipment two or user user b in operator two or operator b so let's start from the left hand side so we you see here that we've got different layers which are colored differently we have the end user functions transport functions transport control functions and service stratum functions so we are going to initiate the request as step number one service is requested from the user equipment it is sent to the service stratum function to the proxy call session control function which in turn actually talks to the serving call session control function which in turn asks the interrogating call session control function to forward or establish a service request process as step number two then the interrogating call session control function may even have to discover or ask or consult a dns server to look for the destination call session control function so the one interrogating call session control function from the calling parties operator network calls the interrogating call session control function entity of the called network of the called party so the destination interrogating call session control function in turn tells the serving call session control function to discover and contact the proxy call session control function which in turn contact the destination user equipment in step number six. So the service request is received. Now if the user presses on the accept button or subscribes to the request then a response is generated. Using the same procedure the response is sent back to the call party. Meanwhile the request is also initiated to look for the appropriate QS availability by the underlying network. For that the resource admission control function is consulted that is going to check if the resources are going to be available. And then correspondingly the mobility management control function is also activated because it's a different network. So the MMCF and RSEF are going to work in collaboration with each other. Then the service response is sent back to the network from which the request was made and then the QS setup request is again made now in the operator one using the same procedure. After doing that then it means that was the last activity of establishing QS at the calling network. And after that the service response is finally sent down to the calling entity that is the user equipment. Now this particular process is finished before actual any data transfer can take place. So this is exactly mentioned in here. We've got a terminal A that sends service request for VoIP call to user B via its own proxy TSCF through its serving CSEF to its interrogating CSEF. But so far everything is in operator A's jurisdiction. Then the operator A interrogating CSEF does DNS look up to obtain the IP address for the interrogating CSEF in the target network. The receiving interrogating CSEF forwards the request for service to the serving call session control function to obtain the contact details of the proxy call session control function because the user is essentially associated with the PCSEF. And after that the service request is sent from A to B which replies and on return each proxy call session control function communicates with the resource and admission control function for admission control. With all this finally concluded we have the resource and admission control function that instructs the underlying network to allocate the bandwidth, the resources, the time, the network attachment because the network attachment actually means the IP address allocation is all going to take place. And then finally we would see that the user data that is the tick line that is voice would be sent from user equipment one in the first operator to the user equipment two in the second operator.