Summary of Core Network Technology (3)
EPC network principle
- EPC network structure and function
There are the following network elements in the EPC network:
eNodeB、MME、HSS、PCRF、SGW、PGW
We can take the process of taking a train as an analogy. If we compare data to passengers, then eNodeB is equivalent to a train station, and data boards from eNodeB. MME and HSS are equivalent to the announcers and administrators of the train station, they tell you which waiting hall to get on the train and where the train is going. When you know where to wait for the train, you need to get on the train through SGW. SGW is equivalent to a ticket gate. After passing this gate, the data will reach the PGW (PDN Gateway) port from the user end, which is equivalent to getting on the train and going to other provinces through here. (Go to other PDN gateways), PGW leads you to the place you want, which is equivalent to the conductor of the train, while PCRF is responsible for handling dynamic policies to manage this process, which is equivalent to the dispatcher.
as the picture shows:
The main network elements on the wireless side, eNodeBs, use the X2 interface, which is a self-learning interface. Only one interface needs to be configured on the eNodeB, and its peer can automatically create a link (equivalent to the existence of a de-link on each eNodeB). dynamic routing to other nodes).
The main network elements on the core side are:
MME, HSS, SGW, PGW, and PCRF, among them, the interface between MME and HSS is S6a interface, and HSS sends user subscription data to MME through this interface, between MME and SGW is S11 interface, between SGW and PGW Between them is the S5 interface, which can be an internal interface or an external interface. In the case of roaming, we call this interface an S8 interface. The PGW is connected to the external network through the SGi interface.
In this figure, the black line and the blue line represent the interface of the control plane and the user plane respectively, that is, the difference between the signaling process and the transmission process.
In the EPC network, according to the number of users and the number of network elements, SGW and PGW can choose two modes of combined configuration and separate configuration. In the combined mode, the s5 interface becomes an internal interface. In the case of separate configuration, the s5 interface becomes an external interface. .
In the case of hybrid networking, the CS network is responsible for voice services, and the 2G/3G/4G LTE network provides data forwarding services to achieve interoperability between 2G and 3G.
There are two types of roaming access methods, the VPLMN (visited network) and HPLMN (local network) they use
- How to access the home network
Signaling process
User access-MME- HSS -SGW- PGW-PCR F-external network forwarding
Among them, the MME and SGW use the network resources of the visited location, and the PGW, PCRF, and HSS use the network resources of the home location.
- How to access the network
Signaling process
User access-MME- HSS -SGW-PGW -PCR F-extranet forwarding
Among them, the MME, SGW, and PGW use the network resources of the visited location, and the PCRF and HSS use the network resources of the home location.
In short, the difference between the access of the visited network and the home network is the PDN gateway of which network is used to enter the access
- EPC network interface and protocol
The Uu interface between the UE and the eNodeB adopts the RRC protocol. The NAS protocol on the UE directly talks to the MME, which is a transparent connection between the eNodeBs. The S1 interface between the eNodeB and the MME uses the S1-AP protocol.
The Uu interface of the user plane adopts the PDCP protocol. When the UE sends data, it will encapsulate the IP packet. After receiving the data, the eNodeB interface will decapsulate it once to obtain the IP data packet. Then use the UDP protocol to re-encapsulate and forward to the SGW. After the SGW receives it, it also decapsulates and encapsulates it once and sends it to the PGW. After the PDN receives it, it decapsulates it again to obtain the user plane message.
S1 interface can be divided into S1-MME interface and S1-U interface
The S1-MME interface uses the S1-AP protocol, the S1-U interface uses the GTP-U protocol, and the GTP-U uses the first version here.
Control plane protocol, using Diameter protocol
The S10 interface uses the GTP-C V2 protocol
Same as S10, also uses GTP-C V2 version
GTP-C V2 is used for signaling control and forwarding GTP-U is used for IP packet forwarding
GX is a mandatory interface, and GXC is an optional interface. When IP protocol is used for transmission between SGW and PGW, GXC can be used to transmit QOS information. When both of them use GTP protocol for transmission, Gxc interface can not be selected.
Diameter is also divided into two types: application layer and base layer.
SGi is the interface from PGW to the outside, and can be transmitted based on RADIUS/DHCP/L2TP.
- Basic process of EPC network
(1) Main process in EPC
Divided into mobility management signaling process and session management signaling process
Among them, the mobile management process mainly includes: attachment, detachment, tracking area update, service request, S1 release, handover, security process (included in other processes as a sub-process)
Session management includes:
Dedicated bearer activation, modification, deactivation, UE requested bearer resource modification, PDN connection establishment, PDN connection release, PDN release process requested by MME
(2) Attachment process
The attach process is used when the UE registers for the first time. A default EPS bearer is established for the first attachment, which provides a permanent IP connection. After that, the user's MM context and EPS bearer context will be created in the MME and UE, and the EPS bearer context will be created in the SGW and PGW. User's EPS bearer context.
Tracking to Update (TAU) can be based on timer TAU procedures associated with timers and MME TAU procedures on MME changes, ie update procedures for those changes based on user time and location.
(3) Connection release
If the eNodeB has not received the context of the client for a long time, in order to save network resources, it will initiate the S1 connection release process. At this time, the connection between UE and eNodeB and MME will be released, and the connection between eNode B and SGW will also be released.
(4) Business request process:
Establish S1-AP connection and S1-U bearer. The triggering reasons are: UE triggering (service triggering) and network triggering (paging, paging request)
(5) Switching process
Handover possible within E-UTRAN (EPS)
Handover based on X2 interface without changing MME, or based on S1 interface handover, MME/SGW may relocate
Can switch between different RATs
LTE—2G/3G/CDMA/WiMax/WLAN
(6) Separation process
The separation process is used in the following situations:
- UE is detached from EPS service
- The UE disconnects from the last PDN
- The network informs the UE that it can no longer access EPS
Separation methods can be divided into explicit separation and implicit separation
Explicit refers to the detachment process that is completed normally, and implicit refers to the detachment process in which the UE voluntarily resets the detachment timer to zero without notifying the network.
4. Basic concept of EPC network
(1) User status management, using EMM (mobility management) and ECM (connection management) to describe user status (registered, unregistered, connected)
(2) MM context and EPS bearer context
MME and HSS will create MM context and EPS bearer context, while SGW and PGW will only save EPS context, that is to say, SGW and PGW do not pay attention to user location information and authorization information
(3) Identification in EPC
Permanent user identity IMSI (international mobile subscribe identity)
It consists of three parts: MCC (Mobile Country Code), MNC (Mobile Network Number) and MSIN (Subscriber Identity).
Temporary user ID GUTI (globally unique temporary ID)
Temporarily identify users from a certain range and a certain period of time
Temporary Subscriber Identity S-TMSI
Composed of MMEC and M-TMSI, it is equivalent to a simplified version of GUTI, which can shorten the length of signaling messages.
(4) Business-related
APN (Access Point Name) access point name
Including NI and OI two parts
APN-NI: defined by operator
APN-OI: defined by 3GPP
- Location related
It is marked with TAI and consists of three parts: MCC, MNC and TAC
- PDN connection
The UE can have multiple PDN connections at the same time, and each PDN connection is uniquely identified by the external PDN network, IP address, PDN gateway, and APN.
- EPS bearer
There are two types of bearers: default bearer and dedicated bearer. The default bearer is created during the attach process and remains active until the user detachs it. A dedicated bearer is a bearer created as needed.
Relationship between PDN connection, EPS bearer and TFT
A user can support multiple PDN connections, their ip addresses and APNs are different, and the user supports up to 11 PDN connections. The bearers of the same PDN project can be divided into default bearers and dedicated bearers, and different bearers are distinguished by different bearer IDs and QoS. The TFT belongs to the bearer, and there are multiple filters in the TFT. The user terminal or PGW matches different service flows to different bearers (according to the bearer ID) through the TFT.
5. GPRS-EPC network architecture
(1), IP bearer service
IP bearer service refers to the function that SGSN supports packet domain to transmit data between users and Packet Data Network (PDN) with IP as the bearer protocol. The main applications of IP bearer services include FTP, Telnet, WWW, E-mail, VOD and other services
(2) Value-added services
Value-added services include CAMEL services, location services, supplementary services, short messages, and cell broadcast services