Table of Contents 1. Introduction 2.1 S1 Protocol Stacks 2.2 S1AP Procedures and Messages Relating to Handover 2.3 S1 Handover Procedure at a Glance 2.4 UE State and Connection Information Before and After S1 handover |
2. Concept of S1 Handover
2.1 S1 Protocol Stacks
S1 handovers are performed between a source eNB and a target eNB through the S1 interface, which connects eNB and EPC. eNB communicates with MME through S1AP signaling in the control plane, and communicates with S-GW through GTP tunnel in the user plane. Figure 1 shows the protocol stacks over the S1 interface in control and user planes.
Figure 1. Protocol Stacks over S1 Interface
When a new eNB is installed, an “S1 Setup” procedure is performed between the eNB and MME(s). The eNB provides the MME(s) with eNB configuration information by sending an S1 Setup Request (eNB ID, eNB Name, TAC) message. Then, each MME returns an S1 Setup Response (GUMMEI, MME Name, Relative MME Capacity) message to the eNB so that it can update its configuration information. Here, Relative MME Capacity parameter is used in load balancing among MMEs in an MME. Its value, marked in weight factor, shows each MME’s relative capacity to handle UE connection. eNBs connected to more than one MME use this value when selecting a MME to establish a new UE connection with.
UE connections between eNB and EPC are as follows: In the control plane, each user’s signaling between eNB and MME is provided through S1 Application Protocol (S1AP) signaling connection1, and identified by {eNB UE S1AP ID, MME UE S1AP ID}. In the user plane, each user’s S1 bearer between eNB and S-GW is provided through GTP (GPRS Tunneling Protocol) tunnel, and identified by {DL S1 TEID (S1 eNB TEID), UL S1 TEID (S1 S-GW TEID)}.
2.2 S1AP Procedures and Messages Relating to Handover
Tables 1 and 2 in 3GPP TS 36.413 S1AP show the elementary procedures of S1AP, which include both non-UE and UE related procedures. This document is about handover, and hence only discusses UE related procedures, particularly those related to the handover procedures to be explained in Chapter III. Table 1 lists the elementary S1AP procedures related to S1 handover, and their associated S1AP messages.
Table 1. S1 Messages for Handover related S1AP Elementary Procedures [2]
Among the S1AP messages listed the table above, those to be used in Chapter III (those used in successful handovers) are briefly explained below.
2.3 S1 Handover Procedure at a Glance
As seen in the previous document [3], an S1 handover procedure consists of preparation, execution and completion phases. Before we go further into detail, we will briefly preview the S1 handover procedure.
Figure 2. Simplified Procedure of S1 Handover
Figure 2 above illustrates at a glance the procedures required before, during (preparation, execution and completion phases) and after S1 handover. For convenience’s sake, S-GW and P-GW are marked as SAE-GW, and source and target eNBs are marked as SeNB and TeNB, respectively.
Before S1 Handover
In the figure above, the UE is being served through eNB A (a serving cell in eNB A, to be more exact) that it has connected to. When the UE detects a measurement event, it sends a Measurement Report message to eNB A.
S1 Handover Preparation
The source eNB (i.e. eNB A in the figure) chooses a target eNB (i.e. eNB B in the figure) to handover to, based on the neighbor cell list information it has kept and the information on the signal strength of the neighbor cells included in the Measurement Report message. Next, it, realizing a handover to the target eNB through the X2 connection is not possible, decides to perform a S1 handover instead, and prepares to perform one through MME. Both eNBs communicate with the MME through S1AP signaling. At this time, the target eNB allocates radio resource in advance to ensure the same services currently provided by the source eNB are also available at the target eNB. The MME also provides the source eNB with the information required for the UE to access the target cell.
In the meantime, the target eNB and S-GW allocate resources needed to create an indirect tunnel through which DL packets arriving at the source eNB are forwarded to the S-GW and finally to the target eNB while a handover is being performed, as follows:
▪ establishes an UL S1 bearer through which to forward UL packets after the handover by using the S1 S-GW TEID obtained from the MME, and allocates S1 target eNB TEID for a DL S1 bearer (❸).
▪ allocates S1 target eNB TEID for the tunnel connecting the S-GW and the target eNB (this tunnel is a part of the indirect tunnel2 connecting all the way from the source eNB, S-GW and the target eNB) to be used for forwarding DL packets while the UE attempts to access (i.e. perform a handover to) the target eNB.
▪ configures a Handover Command message that includes information needed for the UE to access the target cell (e.g. Target C-RNTI, Target DRB ID, etc.).
▪ and sends the information to the MME by including it in a Handover Request Ack message (❹).
▪ Creates an indirect tunnel connecting the target eNB (❻).
▪ allocates S1 S-GW TEID for the tunnel connecting the source eNB and the S-GW (this tunnel is a part of the indirect tunnel connecting all the way from the source eNB, S-GW and the target eNB), and sends it to the MME through a Create Indirect Data Forwarding Tunnel Response message.
Through Steps ❽ and ❻, the entire indirect tunnel connecting all the three entities, the source eNB, S-GW and target eNB, is created.
S1 Handover Execution
Now the two eNBs are ready to perform a handover, it is time to command the UE to perform one.
▪ commands the UE to perform a handover to the target cell by sending a Handover Command message that includes all the information needed for the UE’s access to the target cell (❶).
▪ informs the MME about from which UL/DL packet it should receive/send from/to the UE by sending an eNB Status Transfer message (❷).
▪ sends the DL packets received from the S-GW on to the target eNB through the indirect tunnel connected to the target eNB via the S-GW (❹).
S1 Handover Completion
As the MME already knew that the UE was about to perform a handover, the target eNB, unlike in X2 handover, does not request the MME for path modification. Instead, the target eNB sends the MME a Handover Notify message to indicate the UE has completed the handover once the UE is connected to the target eNB.
▪ It stops DL packet delivery by sending an End Marker (EM) packet through the DL S1 bearer connected to the source eNB.
▪ Then it creates a DL S1 bearer that connects to the target eNB, and resumes DL packet delivery to the target eNB.
▪ It sends DL packets arriving through the indirect tunnel to the UE until an EM packet arrives.
▪ Once an EM packet arrives, it sends the UE the ones arriving through the new path.
▪ requests the source eNB to release S1 resources related to the source eNB and the UE Context it has by sending an UE Context Release Command message (❹).
▪ request the S-GW to release resources associated with the indirect tunnel by sending a Delete Indirect Data Forwarding Tunnel Request message (❺).
After S1 Handover
The UE is now being served through eNB B (the serving cell at eNB B, to be more exact) that it has connected to.
2.4 UE State and Connection Information Before and After S1 Handover
Figure 3 illustrates the connection establishments in the user/control planes, and the UE and MME states before, during and after the S1 handover.
• Before S1 Handover
The UE stays in EMM-Registered and ECM/RRC-Connected and keeps all the resources allocated by E-UTRAN and EPC.
• During S1 Handover
Even during the handover phase, the UE’s state on the NAS layer remains unchanged. Both the source and target eNBs are connected to the MME through the S1 signaling connection established over the S1-MME interface. They are also connected to the S-GW through the indirect tunnel created over the S1-U interface for DL packet forwarding. In Figure 3, Step 2) shows the connections and states while the handover is interrupted during the handover execution phase. During this period, no radio link connection is active, but the UE still remains Connected.
• After S1 Handover
The UE remains in EMM-Registered and ECM/RRC-Connected states. The E-RAB (DRB + S1 bearer) path is switched to connect to a new eNB in the user plane while a new RRC connection is established in the control plane.
Figure 3. Connections and States before/after S1 Handover
It's very good in details. Thank you
Excellenbt details to make everythings clear and easy to understand.
Looking forward to similiar document for CSFB to 3G/2G and VoLTE Call Setup.
So good!But a question,why not use C/D instead B/A? Because C/D+RRC=E-RAB
Thanks for your detail document. Can you publish document relating inter-RAT handover
I am blown away,
this is the best informative website ever. could you tell me which tools do you use to draw your figure?.
Hi Team,
In the handover preparation diagram, it shows ECM and EMM state at Target ENB.
Shouldn's they appear as MME states?
Regards
Gautam
Excellent learning material, thanks. Do you have content for the two other S1 handover procedure especially on Inter-LTE HO to understand why actually that is triggered?
• Inter-LTE Handover: UE’s serving MME and/or S-GW is changed after handover
■ Inter-MME Handover: UE’s serving MME is changed, but S-GW remains unchanged after handover
■ Inter-S-GW Handover: UE’s serving S-GW is changed, but MME remains unchanged after handover
■ Inter-MME/S-GW Handover: Both UE’s serving MME and S-GW are changed after handover
• Inter-RAT Handover: Handover between networks that use different radio access technology
■ UTRAN to E-UTRAN
■ E-UTRAN to UTRAN, etc.