When I turn on a terminal, it establishes a radio connection. If I don’t use my terminal, if I leave it in my pocket, are the connections still maintained? Alternatively, if I want to maintain a radio connection, won’t it consume too much energy? We are going to answer these two questions in this video. At power up, we have already seen the establishment of a default bearer. A packet that arrives from the network must be routed as quickly as possible. Likewise, a packet that is created because I have an app running on my terminal must also be transmitted as quickly as possible. When using my terminal, I can move (as long as I am not driving my car of course) an active radio connection requires signal measurements from different base stations to be sent regularly in order to detect whether the terminal is getting farther from one base station and closer to another. If I’m not using my terminal, maintaining this radio connection requires regular exchanges of measurements and checks to ensure that the terminal is still covered by the same cell. All that would entail the expenditure of a lot of energy. And the batteries would not last long. Furthermore, for each radio connection, an RNTI should be allocated. RNTIs are 16 bits in length, which makes over 65,000 available. 65,000 seems a huge number but it isn’t that large. So, we’re going to avoid maintaining the radio connection all the time. What is implemented is a time-out when the radio connection is inactive. As long as I’m using my terminal, that is to say, as long as user data is being transmitted, the connection is maintained and the inactivity timer is reset. You can see it here. There’s no more activity. The timer is counting down. It’s counting down, but the user, for example, loads a Web page. There is activity once again. The timer is reset. When the value reaches zero, in other words, when no application has sent any message for a defined period of time, the radio connection is released. This timer is placed in the eNodeB and is started at the end of an exchange. At the end of the time out, the RRC connection is released and the UE loses its RNTI. A radio connection is a bit like a leash between a dog and his master. Note that the dog is the UE and the master is the eNodeB. Sometimes, to allow the dog to freely run, the master removes the leash. This is the radio connection release procedure. If we look at the network, we have here a representation of all the connections and tunnels. We have already seen these diagrams in previous sequences. If there is radio activity, the connection and all tunnels are maintained. If there is no activity the radio connection is released. But, if we release the radio connection, that means the network loses awareness of the cell of the UE. So, the terminal can move or change base stations without the network necessarily being aware. We’ll see this in week six. This means that the network no longer knows precisely under which eNodeB the terminal is located. There is no reason to maintain an S1 bearer between the eNodeB and the serving gateway. There is no reason to maintain an S1-AP connection. So, when we release the radio connection we release the S1 bearer and the S1-AP connection at the same time. This is referred to as an ECM state. In the ECM_Idle state, there is no radio connection; the network is unaware of the precise location of the terminal. On the other hand, the terminal remains connected to the network. However, during the ECM_Connected state, the radio connection is active, an RNTI is allocated and there are signaling and user bearers. Bearers and tunnels are always maintained between the SGW, the PGW and the MME. A timer is managed by the eNodeB. In case of inactivity of the UE, the radio bearers, the S1-AP connection and the S1 bearer are released. There are, therefore, frequent switches between the ECM_Connected and ECM_Idle states.
