Dual Active Protocol Stack Handover

Description

DAPS (Dual Active Protocol Stack) Handover is a mobility procedure introduced in 3GPP Rel-16 that maintains the source eNB connection after reception of the RRC message for handover and until releasing the source cell after successful random access to the target eNB. The core idea is that during handover, the UE maintains two active protocol stacks — one for the source cell and one for the target cell — so that user data can continue to be transmitted and received on the source link until the target link is ready, thereby minimizing the data interruption time that is inherent in conventional handover procedures.

In a conventional handover, the UE typically detaches from the source cell, performs synchronization and random access to the target cell, and only then resumes data transfer. This creates a gap in user-plane data, which can be particularly disruptive for latency-sensitive applications such as voice, video, and URLLC services. DAPS Handover addresses this by allowing the UE to keep the source link active while simultaneously establishing the target link. The UE continues to receive downlink data from the source eNB and transmit uplink data to the source eNB until the target link is successfully established, at which point the UE switches data transfer to the target eNB and releases the source connection.

Architecturally, DAPS Handover involves coordination between the source eNB, the target eNB, and the core network. The source eNB decides to initiate a DAPS handover and sends a Handover Request to the target eNB, including an indication that DAPS is requested. The target eNB prepares the necessary resources and returns a Handover Request Acknowledge with a DAPS configuration. The source eNB then sends an RRC Reconfiguration message to the UE containing the target cell configuration and the DAPS indication. Upon receiving this message, the UE does not immediately detach from the source cell; instead, it establishes the target protocol stack while keeping the source stack active. The UE performs random access to the target cell, and upon successful completion, it sends an RRC Reconfiguration Complete message to the target eNB. The target eNB then informs the core network of the path switch, and the source eNB is released.

Key technical challenges addressed by DAPS Handover include the management of two simultaneous protocol stacks in the UE, the handling of security keys for both source and target links, the coordination of data forwarding between source and target eNBs, and the avoidance of data duplication or loss during the transition. The UE must be capable of maintaining two active radio links, which requires dual connectivity-like capabilities. Security is handled by maintaining separate security contexts for the source and target links, with the target security context activated upon successful random access. Data forwarding from the source eNB to the target eNB ensures that in-flight packets are not lost during the switch.

DAPS Handover is specified in 3GPP TS 36.300 (E-UTRA overall description), TS 37.483 (multi-connectivity), TS 38.401 (NG-RAN architecture), and TS 38.463 (E1AP). It is applicable to both LTE and NR, and is particularly beneficial for URLLC and other latency-sensitive services where even brief interruptions in data transfer are unacceptable. The procedure has been maintained through Rel-19, reflecting its importance in 5G and beyond for ensuring seamless mobility.

Purpose & Motivation

DAPS Handover was created to solve the problem of data interruption during handover in mobile networks. In conventional handover procedures, the UE must detach from the source cell before it can fully establish the target cell connection, resulting in a period during which user data cannot be transmitted or received. This interruption, often referred to as the handover interruption time, can range from tens to hundreds of milliseconds depending on the scenario, and it is particularly problematic for latency-sensitive applications such as voice over IP, video streaming, online gaming, and URLLC services in 5G.

Prior to DAPS, techniques such as make-before-break handover and dual connectivity could reduce interruption time, but they had limitations. Make-before-break allows the UE to establish the target link before releasing the source link, but it does not allow simultaneous data transfer on both links. Dual connectivity allows simultaneous data transfer on two links, but it is designed for long-term aggregation rather than handover and involves complex core network anchoring. DAPS Handover combines the benefits of both approaches by allowing simultaneous data transfer on source and target links specifically during the handover execution phase, without requiring long-term dual connectivity anchoring.

The introduction of DAPS in Rel-16 reflects the 5G era's stringent requirements for reliability and low latency. By minimizing handover interruption, DAPS helps ensure that critical services maintain their quality of experience as users move through the network. It also supports the broader 5G goal of seamless mobility across heterogeneous deployments, including LTE and NR interworking scenarios.