LTE Radio Access Network Architecture

The E-UTRAN comes with a simple architecture that is illustrated in Figure 1. The base stations of the network are called eNodeB and each eNB is connected to one or multiple MMEs. These MMEs in turn are connected to a S-GW that may also be co-located (comprising the same physical hardware) with the MME. The interface between the eNB and MME is the called the S1 interface. In case the MME and S-GW are not found in the same physical entity, the S1 control plane interface (S1-MME) will connect the eNB and MME while the S1 user plane interface (here S1-U) will connect the eNB with the S-GW.

 
Figure 1: E-UTRAN network architecture (according to 3GPP 36.300). 

In case one eNB is connected to multiple MMEs, these MMEs form a so-called MME pool and the appropriate network functionality is called S1 flex. The initial signaling procedure used to connect an eNB with a MME is the S1 setup procedure of the S1 Application Part (S1AP).

The X2 interface is used to connect eNBs with each other. The main purpose of this connectivity is intra-E-UTRAN handover. In the real world it will not be possible for all eNBs of the network to be connected via X2 due to limited transport resources on the wired interfaces. It also must be expected that, physically, the X2 links will lead from one eNB to the MME and then back to a second eNB. In other words, the hubs will be located at the physical location of the MME.

It is important to understand that only the base stations and their physical connections (wires or fibers) are defined by 3GPP as the E-UTRAN, while MME and S-GW are seen as elements of the EPC network.