Research on WiFi-Cellular Network Seamless Handover Technology for Mobile Devices

This article discusses the technical challenges and solutions for WiFi to Cellular network handover in smart mobile devices. Currently, mobile terminals typically use only one wireless resource at a time, and when switching between WiFi and Cellular networks during user movement, the link reconstruction causes significant delays, resulting in user experience issues such as stuttering or connection interruptions. The example shows that Zoom application experiences disconnection times exceeding 15 seconds during WiFi to Cellular handover.

The problem is analyzed through a model where three time components are defined: T1 (detection time from WiFi signal degradation to handover decision), T2 (Cellular radio link establishment time), and T3 (TCP connection reconstruction time). The core challenge is minimizing these three time components while balancing power consumption and network costs.

Three key technologies are proposed to address these issues:

1. Dynamic Link Evaluation: Similar to navigation software evaluating routes based on traffic conditions, this approach evaluates WiFi and Cellular link quality using metrics like application latency, RSRP signal strength, and RTT. This enables proactive link switching before complete disconnection, reducing T1 and preparing data paths in advance to reduce T2. However, it cannot completely solve T1=0 due to varying QoS requirements across different applications and challenges in accurately predicting target link conditions.

2. QUIC (Quick UDP Internet Connection): QUIC is Google's UDP-based low-latency internet transport protocol. Traditional TCP connections are identified by a five-tuple (source IP, source port, destination IP, destination port, protocol), so when network interfaces change during WiFi-Cellular handover, the IP change requires TCP connection reconstruction, causing T3 overhead. QUIC solves this through connection migration by using a 64-bit random number as the connection ID instead of IP+port+protocol. As long as the ID remains unchanged, the connection persists even when IP or port changes, allowing T3 to approach zero. However, QUIC cannot solve the T1 problem.

3. MPTCP (Multipath TCP): MPTCP extends TCP to allow simultaneous multiple TCP connections between communication parties, fully utilizing multiple paths for throughput aggregation or reliability improvement. MPTCP operates between Socket and TCP layers, managing multiple TCP subflows. It supports three modes: Aggregation (different subflows transmit different data for high throughput), Redundant (different subflows transmit same data for high reliability), and Backup (subflows in backup mode are used only when no other subflows are available). The Redundant mode is recommended for minimizing T1-T3 delays, though it requires additional traffic and power consumption, plus server-side support.

The article concludes that each solution has its advantages and disadvantages, and the most suitable method should be chosen based on actual requirements. Additionally, with advances in AI technology, combining accurate user trajectory prediction and precise indoor positioning could enable truly seamless handover between different communication systems without additional traffic, power, or network resource costs.