Can you have quality of service in a 4G network? That is the question we are going to answer in this video. To understand the context of this question, we need to look back at what we’ve seen in the course. As soon as the UE powers up, it attaches to the network and, at the same time, a set of bearers is established. These bearers are appended to form what is called the default bearer between the UE and the PGW. From the moment the 4G terminal powers up it has a default bearer at all times. This bearer is not associated with a level of quality of service. Because we have a network based on IP with queues in different routers, there can be fluctuations of the transit delay through time. If a router’s memory is full, it will lose packets and therefore there can be packet loss. The default bearer is always available but it is not associated to a quality of service. That’s fine for a large number of applications. Actually, it is generally the TCP protocol that manages the retransmission of packets. So, if I’m surfing the Web or if I have applications that stream, that’s just fine. We can use a 4G terminal, as we well know, with a large number of services. However, for certain applications such as voice over IP, it is necessary to guarantee low delay in the network. Low latency is necessary for some critical applications as well. The default bearer cannot guarantee that because all packets are processed with the same priority level, which is another way of saying that the concept of priority is absent. So in a 4G network, we’ll define a dedicated bearer that’s going to include packets that will have a TEID different from the TEID of the default bearer. Because of this different TEID, it can be associated to a certain quality of service. In real terms, that means that these packets, transported on dedicated tunnels, will be managed preferentially by the operator compared to the other packets. If we have a queue, for example, the packet will go to the head of the queue. This way, the packets transiting on the dedicated bearers will statistically arrive before the others. This is called the guarantee of quality of service. Regarding, the management of the quality of service, there are mechanisms on IP, like Diffserv for example which is used a lot. The dedicated bearer is in addition to everything else. We can see it on this diagram: we see the Bearers linked to data, we keep a tunnel for control and an S1-AP connection and the dedicated Bearer is added. There can be one or many dedicated bearers. The request is generally made by the cell phone, but it is made using the default bearer and using special signaling protocols. For example, the SIP protocol can be used to establish a voice over IP communication and the network analyzes the SIP protocol messages, and at that moment, it will see that it is necessary to establish a dedicated bearer. Why do we leave the establishment of the dedicated Bearer up to the operator or under the control of the operator? Well, precisely because, if everyone established a dedicated Bearer, it would be difficult to give quality of service. So, behind quality of service, there are admission control mechanisms somewhere. In any case, the operator must be able to control setting up dedicated Bearers to be able to guarantee quality of service. The dedicated Bearer corresponds to the same context as the default bearer, that is, it is linked to the same Access Point Name and the same IP address. That is what we’ll be looking at in the course. It is possible to establish several default Bearers, which is to say that, according to the standard (it’s not necessarily a provided service) a cell phone can request several IP addresses. We will not be dealing with this in the course. In the initial state, we have our now classical configuration. We assume that the terminal is connected to the network. This means that it is in state EMM Registered and that the radio connections and S1-AP are established. In other words, the terminal is in state ECM-Connected. The dedicated bearer activation procedure works as follows: It’s undertaken by the P-Gateway. The P-Gateway sends a GTP Create Bearer Request message and it indicates the QoS level. QoS stands for Quality of Service. The S-Gateway notifies the MME of the request, it also passes on the QoS level and the MME retransmits 2 things in one message. It transmits an S1AP E-RAB Setup Request to indicate to the eNodeB that there will be quality of service to be managed, that is, some packets must have a high priority. So, the eNodeB must be in the loop. There is an impact on the scheduler because the scheduler chooses the radio resource blocks to be allocated to the various terminals. The eNodeB must also notify the terminal, since it also concerns the application on the terminal, that there is bearer with a good QoS level. That is the goal of the ESM Activate Dedicated Bearer Context Request message that we see here. The eNodeB retransmits an indication to the UE that there was a change on the radio connection and, on top of that, there’s a dedicated Bearer. The UE acknowledges when it has taken the modification of the radio configuration into account. This acknowledgement causes the “Setup Response” message to be sent to indicate that the request has been taken into account. So, when the application, for example, the upper layers have taken into account the establishment of the dedicated bearer, the Session Management layer sends a message acknowledging the previous message. It is Activate Dedicated Bearer Context Accept. This is sent to the MME: it’s a NAS, Non Access Stratum, message. And, at that moment, the MME knows that the cell phone has taken the establishment of this dedicated bearer into account, that the eNodeB has also taken the management of quality of service related to the establishment of this dedicated bearer into account. The MME can send the response to the request of the P-Gateway via the S-Gateway. All these messages include the TEIDs that correspond to the extra bearer. At the end of this procedure, we have this established dedicated Bearer. Various TEIDs are used for the various default and dedicated Bearers. You can activate, so you also need to think about deactivation. Deactivation happens the same way, undertaken by the P-Gateway. We won’t go into the details of this exchange of messages. You can look at them on your own after the video. It always goes through the Serving Gateway, the MME, a request to the eNode B and a NAS message to the terminal. When the eNodeB and the MME have taken the release of the dedicated bearer into account, acknowledgement messages are sent all the way up to the P-Gateway. At the end of the procedure, we’re back to square one with just the default bearer. The default bearer is of course maintained; connectivity is maintained.
