Reducing mobility interruption time in 5G networks
Tomorrow’s 5G ultra-reliable low-latency use cases, such as in transport and manufacturing, demand cell handover latency very close to zero milliseconds – a significant reduction on today’s 30-60 millisecond handover in 4G LTE systems. Find out how a new 3GPP solution could make all the difference for 5G networks.
The next steps for 3GPP, release 16 and 17, will introduce additional features to further enhance the support for new use cases related to smart manufacturing, connected vehicles, electrical power distribution and more, such as drones which are controlled by the network.
All of these potentially critical use cases require ultra-reliable low-latency communication (URLLC) which means high reliability and availability, as well as very low end-to-end latency in the range of a few milliseconds. The 5G system is designed with this in mind, and the continued evolution of 5G New Radio (NR) will continue to optimize mobility performance further. A critical part of this is to reduce the handover interruption time between cells in the 5G network.
What is handover interruption time?
At handover from a source cell to a target cell, there is a brief time when the mobile terminal can’t transmit or receive user data. This mobility interruption time can be defined as the shortest time duration supported by a mobile network during handover.
In a 4G LTE deployment, the mobility interruption time is typically around 30 – 60 milliseconds depending on handover scenario and effective radio conditions. However, to ensure the performance of emerging 5G wireless use cases, the 3GPP community has faced a gigantic challenge to shorten this time – well, quite significantly – to as close to zero milliseconds as is technically possible.
In today’s 5G networks, such a short mobility interruption time is possible in a few scenarios, for example when the mobile terminal moves from one beam to another within the same cell.
The solution which will shortly be standardized as part of 3GPP Release 16 will bring shorter interruption times to more handover scenarios and deployments. For us at Ericsson, it has been important to reduce the interruption time for the most common type of handover scenario deployed in the networks today: the intra-frequency handover scenario.
A look at the proposed 3GPP RAN solution
In both today’s 5G networks and 4G legacy networks, the mobile terminal typically releases the connection to the source cell before the link to the target cell is established. That is to say, uplink and downlink transmission is finalized in the source cell before the mobile terminal starts to communicate with the target cell. However, the problem therein is that this inevitability causes an interruption in the range of a few tens of milliseconds in the communication between the network and the mobile terminal.
So how can we overcome this? Well, the principle of the proposed 3GPP solution is that the mobile terminal will allow the connection to the source cell to remain active for reception and transmission of user data, until it is able to send and receive user data in the target cell. This places a new requirement on the mobile terminal to simultaneously receive and transmit data in both the source cell and in the target cell for a short period during the handover procedure.
The requirement is actually reflected in the feature name selected by 3GPP, namely Dual Active Protocol Stack (DAPS) handover.
The technical view: Dual Active Protocol Stack (DAPS) handover
The main characteristics of the reduced mobility interruption solution are:
- Continued transmission/reception in the source cell after receiving the handover request
- Simultaneous reception of user data from source and target cell
- Uplink transmission of user data switched to target cell at completion of random access procedure
Upon receiving the request to perform a handover with reduced interruption time, the mobile terminal continues to send and receive user data in the source cell. At the same time, a new connection to the target cell is established and the mobile terminal performs synchronization and random access in the target cell. The mobile terminal will establish a new user plane protocol stack for the target cell, containing PHY (Physical), MAC (Medium Access Control) and RLC (Radio Link Control) layers, while keeping the source user plane protocol stack active for transmission and reception of user data in the source cell.
Since the mobile terminal will receive user data simultaneously from both the source and target cell, the PDCP (Packet Data Convergence Protocol) layer is reconfigured to a common PDCP entity for the source and target user plane protocol stacks. To secure in-sequence delivery of user data, PDCP Sequence Number (SN) continuation is maintained throughout the handover procedure. For that reason, a common (for source and target) re-ordering and duplication function is provided in the single PDCP entity.
Ciphering/deciphering and header compression/decompression need to be handled separately in the common PDCP entity, depending on the origin/destination of the downlink/uplink data packet.
The target node controlling the target cell must also be prepared to transmit user data at the moment the mobile terminal is ready to receive it in the target cell. For that reason, user data received from the 5G Core is forwarded from the source node controlling the source cell to the target node while, at the same time, the same user data is transmitted to the mobile terminal in the source cell. The forwarded user data is buffered in the target node until downlink transmission is started.
Once the mobile terminal has completed the random access procedure in the target cell, uplink transmission of user data is switched from the source to target cell. The mobile terminal now informs the target node of the last received data packet in the source cell. Based on this information, the target node can avoid sending duplicate downlink data packets to the mobile terminal – packets that the mobile terminal already received in the source cell.
The target node informs the source node of the successful handover which will trigger the source node to stop its downlink transmission to the mobile terminal.
In the figure above, we can also see that the proposed solution does not involve simultaneous uplink transmission of user data. The mobile terminal transmits user data in the source cell until the random access procedure is completed, and thereafter only in the target cell. This simplifies implementation both in the mobile terminal and in the network since duplication check and in-sequence delivery to the 5G Core is always done either in the source node (until successful handover) or in the target node.
Here at Ericsson, we are very satisfied with the proposed technical specifications of the new solution, and look forward to it becoming part of the 5G standard. We believe that it is an important step in our pursuit to continuously improve the performance of 5G networks and possibilities of future technologies.
Find out more about 3GPP release 16 and 17 in this Ericsson Tech Review article on 5G New Radio evolution.
Read more about our plans to test drive 5G URLLC at Audi’s P-Labs facility in Germany.
Go a little further back in time to when LTE systems required handover latency around the 50-millisecond mark in this technical paper on LTE handover design principles and performance.
Find out more about the role and impact of 3GPP standardization.