Table of Contents
This document presents the LTE network architecture as the first technical document of “LTE” area. The LTE network called EPS (Evolved Packet System) is an end-to-end (E2E) all IP network; EPS is divided into two parts - LTE part which deals with the technology related to a radio access network (E-UTRAN) and EPC part which deals with the technology related to a core network. An E2E all IP network means that all traffic flows – from a UE all the way to a PDN which connects to a service entity – are transferred based on IP protocol within EPS.
In order for LTE services to be provided to a user over the LTE network, an E2E LTE network reference model (NRM) is generally comprised of three additional domains - BSS/OSS domain for subscriber, application domain for providing services, and IP transport network domain for sending IP packets – in addition to basic EPS domain. The scope of this document is focused on the basic EPS domain. Other EPS domain features, such as the architectures for LTE interworking with 3GPP/non-3GPP and roaming, are out of the scope of this document and will be described in other technical documents later.
The remainder of this document is organized as follows. In Chapter 2, the LTE network reference model is defined and the EPS entities and interfaces are explained. Chapter 3 will describe the protocol stacks and then Chapter 4 will explain traffic flows over the LTE network.
2. LTE Network Reference Model
Figure 1 shows an LTE network reference model, consisting of LTE entities (UE and eNB) and EPC entities (S-GW, P-GW, MME, HSS, PCRF, SPR, OCS and OFCS). A PDN is an internal or external IP domain of the operator that a UE wants to communicate with, and provides the UE with services such as the Internet or IP Multimedia Subsystem (IMS). In the following, Table 1 and Table 2 show the functions of the LTE and EPC entities. Table 3 lists the reference points of the LTE network reference model and gives a description of interfaces between EPS entities.
Figure 1. LTE network reference model
Table 1. LTE entities
Table 2. EPC entities
Table 3. LTE interfaces
3. LTE Protocol Stacks
Based on the EPS entities and interfaces defined in Chapter II, the LTE protocol stacks for the user plane and control plane are described in Chapter III.
3.1 User plane protocol stacks
Figure 2 shows the user plane protocol stacks for the LTE network reference model shown in Figure 1. The functions of the key layers of the protocol stacks are briefly described below.
Figure 2. LTE user plane protocol stacks
1) LTE-Uu interface
2) S1-U/S5/X2 interface
3.2 Control plane protocol stacks
Figure 3 shows the control plane protocol stacks for the LTE network reference model. The functions of the key layers of the protocol stacks are briefly described below.
Figure 3. LTE control plane protocol stacks
1) LTE-Uu Interface
2) X2 interface
3) S1-MME interface
4) S11/S5/S10 interfaces
5) S6a interface
6) Gx interface
7) Gy interface
8) Gz interface
On S11 interface, does it provide GTP tunnel per bearer?? I think it should be GTP tunnel per user on S11 Interface.
Thank you for noticing the error.
We have fixed it and posted the updated file.
Thank you again for your kine comment and sorry for the late reply.
Could you please provide LTE roaming call establishing e2e doccuments.
Thanks for sharing valuable doccuments on LTE.
email id: email@example.com
In LTE network how many MME is needed? or How many user one MME can support?
Netmanias LTE series would be best of the best to understand LTE network. Thank you alot!
hi, I think that: there are two way for offline charging is OFCS and CG. They use interface Gz (incluce Ga and Rf). If it is OFCS, the interface is Rf and uses Diameter protocol. If it is CG, the interface is Ga and uses GTP' protocol. Please check again!
In Above Diagram it shows Interface b/w OFCS and P-GW is Gz which will use GTP Protocol .. but Gz is actually uses diameter protocol isn't it ?? got confused can u pls explain it clearly
Hi, Nguyen & Rajeswari,
As far as I know OFCS can have both Gz (GTP') and Rf (Diameter) interfaces as shown figure below.
If my understading is correct, the Netmanias document (figure 3 above) is not wrong.
I don't understand this paragraph
"When a GTP tunnel is established on the S1-U interface, the S-GW assigns a TEID (UL S1-TEID in Figure 4(a)) for uplink traffic and the eNB assigns a TEID (DL S1-TEID in Figure 4(b)) for downlink traffic."
The eNB encapsulates the user IP packets with the S1 GTP tunnel header and forwards the resulting outer IP packets to the S-GW.
So why S-GW assigns TEID for uplink traffic? Why not eNB assigns??
The S-GW, based on the value "UL S1-TEID" in the received packet from the eNB, figures out to which P-GW and with what "UL S5-TEID" to send the packet. If eNB assings "UL S1-TEID", there's no way to know mapping between "UL S1-TEID" and "UL S5-TEID" in S-GW.
That is, packet receving node should assigns TEID as shown below:
- UL S1-TEID is assinged by S-GW during attach procedure, then eNB uses UL S1-TEID when it sends packet to S-GW
- UL S5-TEID is assinged by P-GW during attach procedure, then S-GW uses UL S5-TEID when it sends packet to P-GW
- DL S5-TEID is assinged by S-GW during attach procedure, then P-GW uses DL S5-TEID when it sends packet to S-GW
- DL S1-TEID is assinged by eNB during attach procedure, then S-GW uses DL S1-TEID when it sends packet to eNB
For more information you can refer to the link below:
This is srilakshmi.
could you please share a document on LTE roaming and how does charing works for roaming in lte.
This introduction was well done. Please help us with more insight with topic like DRX, Power, scheduling and others more.