TL;DR
This paper tackles the challenge of signal blockage in millimeter-wave (mmWave) backhaul networks, which is crucial for dense small cell deployments in 5G. They propose a novel Relay-Assisted and QoS Aware Scheduling (RAQS) scheme that intelligently selects relays for blocked flows and optimizes transmission scheduling to maximize the number of flows meeting their Quality of Service (QoS) requirements.
What Problem Does This Paper Address?
The rapid growth of mobile data demand necessitates dense deployments of small cells to increase network capacity. Millimeter-wave (mmWave) frequencies offer abundant bandwidth for backhaul links, enabling multi-gigabit transmission rates. However, mmWave signals are highly susceptible to blockage by physical obstacles, which can severely disrupt data transmission and impact Quality of Service (QoS), especially for delay-sensitive applications. Existing solutions often overlook this blockage issue or don't adequately consider QoS requirements. This paper aims to address this by ensuring that blocked flows can still be reliably served while meeting their QoS demands, thereby maintaining overall network performance.
Methods
The proposed RAQS scheme consists of two main components:
- Relay Selection Algorithm: For blocked flows, RAQS identifies potential relay nodes. It first narrows down candidate relays to those within a specific geometric region around the source and destination. Then, it applies a criterion based on the time ratio (TR) of backhaul transmission to relay path transmission to select relays that offer high rates. To avoid contention, it prioritizes assigning unique relays to different blocked flows. If multiple flows select the same relay, a heuristic approach is used to assign the relay to the flow that needs it most, potentially assigning a suboptimal relay to others or marking them as unservable if no suitable relay is found.
- Transmission Scheduling Algorithm: This algorithm handles both backhaul and relay paths concurrently. It formulates the problem as a mixed-integer nonlinear programming (MINLP) problem, aiming to maximize the number of completed flows. Due to the complexity of MINLP, a heuristic approach is employed. This involves constructing a contention graph to identify interfering links and then scheduling flows slot by slot. The algorithm prioritizes flows that are closer to completing their QoS requirements or have fewer contention issues, ensuring efficient use of limited time slots and maximizing the number of flows that meet their QoS.
Datasets
Not discussed in the paper. The evaluation is based on extensive simulations with randomly generated network topologies and flow parameters.
Key Insights and Ideas
- Relays as a Solution to Blockage: The paper highlights that strategically deployed relays can effectively circumvent mmWave signal blockage, providing a viable alternative path for data transmission.
- QoS-Centric Relay Selection: Simply selecting any available relay is insufficient. RAQS emphasizes selecting relays that offer high link rates and are distinct for different blocked flows to maximize the chances of meeting QoS requirements within the limited time.
- Concurrent Transmission Optimization: The scheme effectively integrates relay paths and backhaul paths into a unified scheduling framework, leveraging concurrent transmissions to boost network efficiency.
- Trade-off in Interference Threshold: The performance is sensitive to the interference threshold (σ). A threshold that is too low can lead to unnecessary contention, while one that is too high can result in inefficient transmissions. An optimal value exists that balances these factors.
- Importance of Relay Selection Parameter (β): The parameter β plays a crucial role in balancing the selection of high-rate relays with the complexity of finding them. A smaller β generally leads to better performance but increases complexity.
Key Results
- Overcoming Blockage: RAQS effectively overcomes the blockage problem, significantly increasing the number of completed flows and overall system throughput compared to schemes that do not address blockage (MQIS, STDMA) or use random relay selection.
- Performance Improvement: In simulations, RAQS demonstrated improvements of up to 37.0% in completed flows and 47.8% in system throughput when all 10 flows were blocked, compared to random relay selection.
- Optimal Interference Threshold: The simulations suggest an optimal interference threshold (σ) around 10⁻², which balances contention and interference for maximizing performance.
- Impact of β: The relay selection parameter β has a noticeable impact, especially with a higher number of blocked flows. A value around 0.53 was found to be a good compromise between performance and complexity.
Limitations and Future Work
- Half-Duplex Assumption: The current model assumes half-duplex nodes, meaning nodes cannot transmit and receive simultaneously. This limits concurrent transmissions.
- Focus on Throughput: The QoS requirement is primarily defined by minimum throughput. Future work could incorporate other QoS parameters like delay.
- Blockage Persistence: The paper assumes blockage is persistent. Investigating dynamic blockage scenarios could be a future direction.
- Full Duplex Technology: The authors suggest exploring the utilization of full duplex technology in mmWave bands to further improve network performance.
- Delay Performance: The paper mentions investigating the delay performance of the proposed scheme in future work.
Summary updated