[Paper Review] Resource Allocation in Cloud Radio Access Networks with Device-to-Device Communications

Yitao Mo, Mugen Peng+3 more|arXiv (Cornell University)|Sep 14, 2017

Advanced MIMO Systems Optimization27 references3 citations

TL;DR

This paper proposes a joint mode selection and resource allocation (JMSRA) algorithm for cloud radio access networks (C-RANs) with device-to-device (D2D) communications to maximize throughput under dynamic traffic and time-varying channels. Using Lyapunov optimization, the stochastic problem is transformed into a delay-aware mixed-integer nonlinear program, decomposed into mode selection, beamforming, and power control subproblems, solved via a modified branch-and-bound and weighted minimum mean square error method, achieving high throughput, low latency, and fronthaul load reduction with a flexible throughput-delay tradeoff.

ABSTRACT

To alleviate the burdens on the fronthaul and reduce the transmit latency, the device-to-device (D2D) communication is presented in cloud radio access networks (C-RANs). Considering dynamic traffic arrivals and time-varying channel conditions, the resource allocation in C-RANs with D2D is formulated into a stochastic optimization problem, which is aimed at maximizing the overall throughput subject to network stability, interference, and fronthaul capacity constraints. Leveraging on the Lyapunov optimization technique, the stochastic optimization problem is transformed into a delay-aware optimization problem, which is a mixed-integer nonlinear programming problem and can be decomposed into three subproblems: mode selection, uplink beamforming design, and power control. An optimization solution that consists of a modified branch and bound method as well as a weighted minimum mean square error approach has been developed to obtain the close-to-optimal solution. Simulation results validate that the D2D can improve throughput, decrease latency, and alleviate the burdens of the constrained fronthaul in C-RANs. Furthermore, an average throughput-delay tradeoff can be achieved by the proposed solution.

Motivation & Objective

• To address the capacity-limited and delay-constrained fronthaul in C-RANs, which limits spectral and energy efficiency.
• To integrate D2D communications into C-RANs to offload uplink traffic and reduce fronthaul burden.
• To design a resource allocation scheme that jointly optimizes mode selection, beamforming, and power control under dynamic traffic and time-varying channels.
• To ensure network stability and meet quality-of-service requirements by incorporating queue state information (QSI) alongside channel state information (CSI).
• To achieve a flexible tradeoff between average throughput and delay in a stochastic, real-time environment.

Proposed method

• Formulates the resource allocation problem as a stochastic optimization problem maximizing overall throughput under fronthaul capacity, interference, and stability constraints.
• Applies Lyapunov optimization to transform the stochastic problem into a delay-aware optimization problem, enabling joint consideration of queue and channel states.
• Decomposes the problem into three subproblems: mode selection (D2D vs. cellular), uplink beamforming design, and power control.
• Develops a joint mode selection and resource allocation (JMSRA) algorithm using a modified branch-and-bound method for mode selection and weighted minimum mean square error (WMMSE) for beamforming and power control.
• Solves the subproblems iteratively to converge to a near-optimal solution that balances throughput and delay.
• Uses a virtual queue framework to enforce network stability and ensure queueing delay requirements are met.

Experimental results

Research questions

• RQ1 How can D2D communications be effectively integrated into C-RANs to reduce fronthaul load and end-to-end latency?
• RQ2 What is the optimal tradeoff between system throughput and delay in a dynamic, time-varying C-RAN with D2D links?
• RQ3 How can resource allocation jointly optimize mode selection, beamforming, and power control under stochastic traffic and channel conditions?
• RQ4 To what extent can the proposed algorithm improve spectral efficiency and fronthaul utilization compared to conventional C-RAN-only or D2D-only schemes?
• RQ5 Can the proposed algorithm achieve stable network operation while maintaining low delay for delay-sensitive applications?

Key findings

• The proposed JMSRA algorithm achieves higher overall average throughput than both C-RAN-only and D2D-only schemes, demonstrating the benefits of D2D offloading in C-RANs.
• The algorithm significantly reduces end-to-end transmission delay by leveraging short-range D2D links for local communication, especially when D2D pairs are close.
• Fronthaul load is substantially alleviated, as D2D communications reduce the amount of uplink data transmitted over the constrained fronthaul links.
• The performance gap between JMSRA and the C-RAN-only mode increases with higher fronthaul capacity, indicating that the gains from D2D are most pronounced under capacity constraints.
• As the maximum D2D link distance increases, the overall throughput of all schemes declines, but the JMSRA algorithm maintains a larger performance gap over the D2D-only mode due to the combined benefits of CoMP and D2D.
• The algorithm enables a flexible average throughput-delay tradeoff, allowing network operators to tune performance based on application requirements.