It can happen, when I’m in the car that I use my terminal and, for example, watch a video. It’s very likely that, during this drive, I will change eNodeBs. How can I continue to watch the video without an interruption of service while my terminal is moving? That is the question we are going to answer in this video. First, we have to take a look back at the concept of cells. We know that a 4G network is made up of a group of eNodeBs that are spread over the territory. Each eNodeB covers a cell. We consider a terminal that has a radio connection with an eNodeB. If the terminal is close to the base station, the signal received from the base station is strong and the data bit rate is high. The more the user distances himself from the base station, the weaker the signal received from the base station becomes and the lower the throughput is. If the signal becomes too weak, well, there’s a risk of losing the radio connection because the data transmitted will be incorrectly received. The radio connection must be transferred from the current eNodeB to the new eNodeB at the right moment. This is called the handover or the handoff. So, the handover is the transfer of a connection,we’re assuming an active connection, from one eNodeB to another eNodeB. Determining the right eNodeB to which we will transfer this connection could be a problem So, we need a mechanism that will enable us to identify each eNodeB on the radio interface. This is one of the functions of the beacon. Each base station broadcasts a reference signal that can be detected by any terminal. We won’t go into the details, but this reference signal is specific to each eNodeB and is linked to what is called the PCI or Physical Cell Identity. The important point is that the terminal is capable of measuring the strength of the signal it receives from each base station. If we assume that the terminal, moves like this. The strength of the signal it receives from base station 1 decreases while the strength of the signal from base station 2 increases. We see here that the strength of the signal from 3 starts to increase and then finally diminishes. What the terminal does is to measure the power levels of the signals received from the six best neighbors. In general, this is a typical value. But these measurements are made by the terminal. The choice made in cellular networks is to leave the important decisions up to the network, since it’s the operator which controls the radio resources. The network has a global view of the load of the system. It has a better position than the UE to determine the best moment to trigger the handover. The terminal will measure the signals but will transfer the measurements regularly to the network. This is called the UE-assisted Network-triggered Handover. In other words, a cell transfer decided by the network but with the assistance of the terminal. This assistance is simply a regular uploading of measurements by the terminal. We’ve depicted it here with these dotted lines. Uploading these measurements enables the eNodeB to determine the target cell. A small note: if the terminal is very close to its eNodeB, it will receive a very strong signal and it won’t be necessary to upload the measurements because there is no need for a handover. So we can imagine that the transfer of measurements is based on a signal threshold. The measurements are reported to the network only if the signal received is lower than this threshold. Now let’s look at what happens on the network side. We have the transfer of the radio connection but not just that. Between the eNodeBs, S‑Gateway, and MME, we have tunnels or connections. The S1 Bearer between the eNodeB and the S‑Gateway will have to be transferred as well at the S1‑AP connection. But that’s not the only thing. If we remember that the user is for example watching a video, the server is sending packets containing the video stream. When the person moves,, he distances himself from the base station and the throughput drops. This means that, in all likelihood, the eNodeB will have to buffer data packets before being able to transmit them on the radio interface. We have, in every node, a buffer. And the most important buffer is in the eNodeB, because it is there, on the radio interface, that we will have the bottleneck. As the terminal gets farther from the eNodeB, the number of queued packets grows. We have to avoid losing those packets. Note that the handover is a complex operation, since we must transfer the radio connection and transfer the tunnels. This operation cannot be done too soon nor too late and the target eNodeB must have enough radio resource to accommodate the terminal. If the cell is saturated and there isn’t enough throughput available because a lot of people are communicating, it might not be the best decision to accommodate a new terminal. These are complex choices that are made by the operators. If we look at what is defined in the standard, there are two types of handovers. The first handover uses the X2 interface, which is defined between two eNodeBs. This is a novel interface defined for 4G networks. When X2 is used, the handover is logically called the X2 handover. In this case, the queue of data packets is transferred from the source eNodeB, that’s to say the eNodeB that is managing the terminal before the handover, to the target eNodeB. So there is practically no break in the connection and no packet loss since the queued packets are transferred from the source eNode B to the target eNode B. The other handover is the S1 handover, for example if there is no possibility to have an X2 interface at that moment, the S1 Bearer is modified and there are mechanisms to transfer the queued packets but they are more complex and therefore there is more of a risk of a temporary loss of packets. We will look at these various handovers in the next videos. To synthesize a bit: for the handover, there is a preliminary phase during which measurements are uploaded. This phase is very important but is before the handover. Next, there’s a preparation phase. Once the decision to make a handover has been taken inside the network, the network will prepare everything necessary to accommodate the UE in the new cell. This is called Handover Preparation. Next, there’s the execution phase, Handover Execution. The eNodeB sends the order to change cells to the terminal. It also reroutes the packets, modifies the tunnels and connections leading to the target eNodeB. And once the terminal has been picked up by the new cell, at that moment, the resources of the former cell are released, this is the final phase of the handover, or Handover Completion.
