Cluster DFS: Building an Integrated ToB/ToC Service Network

As 5G empowers a wide range of industries, its user base has expanded from individual consumers to enterprise customers across sectors like healthcare, education, transportation, energy, and industrial manufacturing. Effectively and flexibly utilizing wireless network resources to meet the diverse needs of these industries has become a key focus of 5G network research.

Cluster DFS: Enabling Fast, Automatic UL/DL Frame Structure Adjustments

Based on real-world commercial network scenarios, different environments exhibit varying data traffic demands. For example, live sports broadcasts primarily rely on real-time data collection, requiring high uplink bandwidth and low latency during live streaming. This necessitates an uplink-heavy frame structure. Once the live broadcast ends, the demand shifts to a downlink-heavy structure for content consumption. In the industrial manufacturing sector, fully connected factories may experience varying demands depending on the area and time, requiring rapid adjustments to the uplink/downlink frame structure based on the specific needs of 5G industrial applications.

Manually adjusting network planning and allocating time-frequency resources based on user and application demands would require significant human and material resources. Additionally, resource allocation strategies must be fine-tuned to avoid performance degradation, and adjusting frame structures may introduce interference due to the coexistence of differing frame structures.

To address these challenges, ZTE has introduced the cluster dynamic frame sharing (Cluster DFS) solution, which optimizes resource allocation strategies to enhance resource utilization while ensuring a high-quality user experience.

Key Technologies of Cluster DFS

Cluster DFS enables a single 5G network to deliver diversified services, meeting both business (ToB) and consumer (ToC) demands (Fig. 1). The key technologies of Cluster DFS include intelligent traffic prediction, cluster formation, frame structure self-adaptation, and cross-cluster interference coordination.

Intelligent Traffic Prediction

Cluster DFS leverages intelligent uplink/downlink traffic prediction to assess and forecast service demands. Different methods are applied based on the application scenario. For typical commercial scenarios, future traffic demands can be predicted using historical data. When traffic demand exceeds a preset threshold, timeslot resources are adjusted accordingly. This method requires dynamic detection of traffic changes and is better suited for scenarios with gradual changes in traffic demand. In scenarios where uplink and downlink traffic follows a predictable pattern, frame timeslot resources can be adjusted based on specific time periods to align with the expected demand.

Cluster Formation

Based on traffic prediction results, sites with similar service demand models are grouped into clusters. The clustering process also considers the geographical distribution of base stations, forming logical clusters that can be continuously updated and refined based on actual performance.

Frame Structure Self-Adaption

Within a cluster, service demands are aligned, allowing for unified frame structure adjustments to meet user needs and ensure a consistent experience. Adjustments are triggered by traffic prediction outputs, which recommend corresponding frame structure changes.

Cross-Cluster Interference Coordination

A critical challenge is addressing cross-cluster interference caused by differing frame structures. Cross-link interference between clusters with different frame configurations can severely impact performance. This issue has been acknowledged in 3GPP’s 5G-A study on TDD flexible frame structures. To mitigate this, effective interference suppression and elimination techniques are required to maintain user experience in both interfered and interfering stations. Key technologies we propose include:

Slot adaptive modulation and coding (AMC): This technology allows different slots to use different modulation and coding schemes, dynamically adjusting to interference levels. Base stations can adapt transmission rates based on slot position, interference strength, and channel quality, lowering the bitrate in interfered slots to minimize the impact on the air interface.

Enhanced interference rejection combining (eIRC): This technique uses SuperMIMO technology to combine signals from multiple distributed antennas, improving interference rejection through enhanced diversity gain. By leveraging the strong correlation between inter-cell interference, eIRC improves channel estimation accuracy and combines received signals to suppress interference, optimizing system performance.

Coordinated beamforming (CBF): Using the spatial freedom of the interfering base station’s antenna array, CBF dynamically adjusts the transmission beam to create spatial nulls in the direction of interference, effectively suppressing interference.

Cluster DFS in Practice

Cluster DFS has been successfully applied in several scenarios, achieving the expected pre-commercial results.

Live event streaming: Cluster DFS played a key role in ensuring communication during the Asian Games, supporting high-bandwidth uplink services required for live streaming and video surveillance. Single-user uplink speeds exceeded 500 Mbps, and overall uplink throughput increased by more than 60% within the venue.

Smart manufacturing: Cluster DFS has been applied in 5G fully connected factories, particularly within production line areas, increasing uplink capacity by more than 70%. It has met the high uplink traffic demands of industrial environments while also providing flexible uplink/downlink capacity for warehouses and handling burst traffic during emergencies.

Commercial square validation: In a commercial test at Kaide Square in Chengdu, Cluster DFS demonstrated an average uplink speed improvement of 67.78%, with peak speeds reaching 556 Mbps. CBF achieved a noise reduction of approximately 8 dB. Cluster DFS is now being deployed in more high-value scenarios, such as concerts and the Chinese Super League.

As ToB and ToC services continue to integrate and industries undergo digital transformation, the demand for spectrum flexibility and resource efficiency is growing. Cluster DFS has proven to be a viable commercial solution, and ZTE will continue to expand its commercial deployment while optimizing performance to enhance support for commercial networks.