One of the main goals of LTE is to provide Self-Optimizing Networks (SON). SON is crucial for network operators because it helps maximize network performance cost-effectively, especially as radio propagation environments change. From the very start, SON has been a fundamental aspect, shaping the design of X2 and S1 procedures.
The S1 interface connects the eNodeB (Evolved Node B) to the EPC (Evolved Packet Core) and is divided into two parts: the control plane and the user plane. The control plane manages signaling and control information, while the user plane handles user data traffic. Below, we’ll look at the protocol structure and functionality of the S1 interface in more detail.
The S1 control plane interface (S1-MME) uses the SCTP/IP stack to ensure reliable delivery of signaling messages. SCTP includes advanced features like multistreaming and multihoming, which enhance reliability and efficiency. The user plane interface (S1-U) operates over a simpler IP-based stack, focusing on efficient data transfer. This split and protocol design help LTE networks achieve high performance and adaptability.
Protocol Structure Over S1.
The S1 protocol structure in LTE relies on a full IP transport stack, eliminating the dependency on legacy SS7 network configurations used in older GSM or UMTS networks. This simplification reduces operational costs when deploying LTE networks.
1. Control Plane:
The S1 control plane utilizes the Stream Control Transmission Protocol over IP (SCTP/IP) stack. Here’s a breakdown of its key features and advantages:
- Reliable Delivery: SCTP ensures reliable delivery of signaling messages, similar to TCP.
- Advanced Features: SCTP includes advanced features such as handling multiple streams to avoid head-of-line blocking and support for multihoming, enhancing transport network redundancy.
- Simplified Protocol Stack: Unlike UMTS, which uses an intermediate connection management protocol, LTE directly maps the S1 Application Protocol (S1-AP) onto SCTP. This direct mapping simplifies the protocol stack.
- Multiplexing: SCTP supports multiplexing, where each stream of an SCTP association handles signaling traffic for multiple individual connections.
Flexibility in Lower Layer Protocols:
LTE offers flexibility in choosing lower layer protocols, allowing operators to select the appropriate IP version and data link layer based on their deployment scenario. For example, operators can start with IPv4 and tailor the data link layer to fit their network needs.
2. User Plane.
The S1 user plane protocol structure is illustrated in Figure below. It is based on the GTP/UDP/IP stack, which is well known from UMTS networks. Using the GTP-User plane (GTP-U) has several advantages, such as the ability to identify tunnels easily and support intra-3GPP mobility.
Similar to the control plane stack, the IP version number and the data link layer in the user plane are fully optional. A transport bearer is identified by GTP tunnel endpoints and IP addresses, including source and destination Tunnelling End IDs (TEID) and source and destination IP addresses.
In practice, the S-GW sends downlink packets for a specific bearer to the eNodeB IP address received in the S1-AP message associated with that bearer. Conversely, the eNodeB sends uplink packets to the EPC IP address received in the S1-AP message associated with the same bearer.
To manage QoS differentiation between bearers, vendor-specific traffic categories (e.g., real-time traffic) can be mapped onto Differentiated Services (Diffserv) code points (e.g., expedited forwarding) through network O&M (Operation and Maintenance) configuration. This ensures that different types of traffic receive appropriate levels of service quality.
Initiation Over S1 Interface.
The S1-MME control plane interface initialization begins with identifying the MMEs that the eNodeB must connect to and setting up the Transport Network Layer (TNL). With the S1-flex function in LTE, an eNodeB must initiate an S1 interface towards each MME node in its pool area. This list of MME nodes, along with initial remote IP addresses, can be pre-configured in the eNodeB during deployment. The eNodeB then initiates the TNL establishment using these IP addresses. Only one SCTP association is established between an eNodeB and an MME.
During the SCTP association establishment, the two nodes negotiate the maximum number of streams to be used over the association. Multiple pairs of streams (each stream is unidirectional) are typically used to avoid head-of-line blocking issues. Among these streams, one pair is reserved for signaling common procedures, while the other streams handle dedicated procedures specific to individual UEs.
Once the TNL is established, basic application-level configuration data is automatically exchanged between the eNodeB and the MME through an ‘S1 SETUP’ procedure initiated by the eNodeB. This procedure is an example of a network self-configuration process in LTE, reducing the configuration effort for network operators compared to manual configurations in earlier systems.
For instance, tracking area identities, essential for system operations as they correspond to zones where UEs are paged, are automatically sent to all relevant MME nodes within the pool area in the S1 SETUP REQUEST message. Similarly, the broadcast list of PLMNs, used when a network is shared by multiple operators, is also automatically configured. This automation reduces configuration effort, minimizes the risk of human error, and ensures alignment between E-UTRAN and EPC configurations regarding tracking areas and PLMNs.
Once the S1 SETUP procedure is completed, the S1 interface becomes operational.
Context Management Over S1 Interface.
In each pool area, a UE (User Equipment) is connected to a specific MME (Mobility Management Entity) for all its communications. This creates a context for the UE in the MME. The first eNodeB (evolved Node B) that the UE connects to in the pool area selects this MME using the NAS Node Selection Function (NNSF).
Whenever the UE becomes active (transitioning from idle to active mode) under an eNodeB’s coverage in the pool area, the MME sends the UE’s context information to that eNodeB using an ‘INITIAL CONTEXT SETUP REQUEST’ message. This allows the eNodeB to create a context and manage the UE while it is active.
Although bearer setup is typically handled by a dedicated ‘Bearer Management’ procedure, the ‘INITIAL CONTEXT SETUP’ procedure also includes the creation of one or more bearers, including the default bearer.
When the UE transitions back to idle mode, the MME sends a ‘UE CONTEXT RELEASE’ message to the eNodeB, which then erases its context for the UE. Only the MME context remains until the UE becomes active again.
Bearer Management Over S1 Interface.
LTE manages bearers through dedicated procedures for setup, modification, and release.
- Bearer Setup: When a bearer needs to be set up, the transport layer address and tunnel endpoint are sent to the eNodeB in a ‘BEARER SETUP REQUEST’ message. This informs the eNodeB where to send uplink user plane data to the S-GW (Serving Gateway). The eNodeB then sends a ‘BEARER SETUP RESPONSE’ message indicating where the S-GW should send downlink user plane data.
- Quality of Service (QoS) Parameters: The QoS parameters for each bearer are also included in these messages, detailing the requirements for that bearer. Standardized QCI (QoS Class Identifier) values are typically used, but operators and vendors can agree on proprietary labels for new services.
This structure ensures that both the uplink and downlink paths are properly established and managed, maintaining the required QoS for each bearer.
Paging Over S1 Interface.
To re-establish a connection with a UE (User Equipment) in idle mode, the MME (Mobility Management Entity) sends a paging request to the relevant eNodeBs based on the tracking areas where the UE might be located. When an eNodeB receives a ‘PAGING REQUEST’ message, it sends a page over the radio interface to the cells within the specified tracking areas.
Typically, the UE is paged using its SAE-Temporary Mobile Subscriber Identity (STMSI). The ‘PAGING REQUEST’ message also includes a UE identity index value, which helps the eNodeB determine the paging occasions when the UE will activate its receiver to check for paging messages.
Mobility Over S1 Interfaces.
1. Intra-LTE Mobility.
In LTE, there are two types of handover procedures for UEs in active mode: the S1-handover and the X2-handover.
X2-Handover
- Primary Procedure: Used for inter-eNodeB handover.
- Interface: Utilizes the X2 interface between eNodeBs.
S1-Handover
- Alternative Procedure: Used when there is no X2 interface or when the source eNodeB is configured to use the S1 interface for handover.
- Design Similarity: Designed similarly to the UMTS SRNS relocation procedure.
- Phases:
- Preparation Phase: Resources are prepared at the target side with the core network’s involvement (steps 2 to 8).
- Execution Phase: The actual handover takes place (steps 8 to 12).
- Completion Phase: Final steps to confirm handover completion (after step 13).
Key Differences from UMTS
- Status Transfer Message:
- Sent by the source eNodeB to carry PDCP status information needed at the target eNodeB (steps 10 and 11).
- Ensures consistent handling of handover by the target eNodeB regardless of whether an S1 or X2 handover is used.
Data Forwarding
- Initiation: Starts after the source eNodeB receives the ‘HANDOVER COMMAND’ from the source MME.
- Types:
- Direct Data Forwarding: Direct path available between source and target eNodeBs.
- Indirect Data Forwarding: When no direct path is available.
Path Switch Update
- HANDOVER NOTIFY Message: Sent by the target eNodeB to confirm the UE’s arrival.
- MME Role: Forwards the message to trigger the path switch update in the S-GW towards the target eNodeB.
- Resource Release: Resources at the source side are released upon reception of a ‘RELEASE RESOURCE’ message from the source MME (step 17).
2. Inter-Radio Access Technologies (RAT) Mobility.
LTE is engineered to facilitate smooth transitions between LTE and other radio access technologies (RATs) like UMTS and CDMA2000, ensuring uninterrupted connectivity for mobile users as they move across different network environments. When transitioning from LTE to UMTS, LTE employs the S1-handover procedures to manage the handover process. Unlike in LTE-to-LTE handovers, there is no need for the STATUS TRANSFER message during the handover steps 10 and 11 because PDCP context continuity between LTE and UMTS is not maintained.
On the other hand, LTE’s mobility towards CDMA2000 involves dedicated procedures designed specifically for this purpose. These procedures focus on tunneling CDMA2000 signaling between the UE and the CDMA2000 system through the S1 interface. This approach ensures that CDMA2000 messages are transmitted without interpretation by the eNodeB, enhancing efficiency and compatibility between LTE and CDMA2000 networks.
For example, the UPLINK S1 CDMA2000 TUNNELLING Message, as illustrated in Figure below, includes crucial details such as the RAT type. This information helps identify the specific CDMA2000 RAT associated with the tunneled CDMA2000 message, ensuring that it reaches the correct destination within the CDMA2000 network infrastructure.
LTE’s architecture and protocols are designed with careful consideration for coexistence with existing technologies. This strategy not only supports seamless mobility but also promotes interoperability across different network standards. By integrating tailored procedures for each RAT, LTE enables mobile operators to deliver reliable and consistent service across diverse network environments, meeting the dynamic connectivity needs of modern mobile users effectively.
Load Management Over S1 Interface.
Load management over the S1 interface in LTE involves three distinct procedures: load balancing, overload handling, and load rebalancing, each designed to optimize network performance under varying traffic conditions.
Firstly, the load balancing procedure aims to evenly distribute traffic among multiple MMEs within a pool area. This process utilizes the Network Node Selection Function (NNSF) integrated into each eNodeB as part of the S1-flex function. By assigning suitable weight factors corresponding to the capacity of each MME node beforehand, eNodeBs statistically balance the load distribution across MMEs without needing further intervention. However, adjustments may be necessary when new MME nodes are introduced or removed, requiring temporary changes to weight factors to manage initial traffic allocation until equilibrium is reached.
During periods of unexpected peak loads, the overload procedure comes into play. When an MME experiences excessive traffic, it can send an OVERLOAD message over the S1 interface to eNodeBs. This message prompts eNodeBs to temporarily restrict specific types of traffic to alleviate the overload condition. The MME controls the extent of traffic reduction by specifying the number of eNodeBs to which the OVERLOAD message is sent and identifying the types of traffic subject to restriction.
Lastly, the load rebalancing procedure enables an MME to offload part or all of its user equipment (UE) to other MMEs swiftly. This function is initiated by the MME using a specific ’cause value’ in the UE Release Command S1 message, directing UEs to reattach to alternative MMEs. Initially targeting idle mode UEs, the procedure can extend to UEs in connected mode if complete MME offload is necessary, such as for maintenance purposes. By executing these procedures, LTE networks optimize resource utilization and maintain service quality even during fluctuations in network demand.