[Paper Review] Hybrid Scheduling in Heterogeneous Half- and Full-Duplex Wireless Networks

Tingjun Chen, Jelena Diakonikolas+2 more|arXiv (Cornell University)|Jan 3, 2018

Full-Duplex Wireless Communications28 references3 citations

TL;DR

This paper proposes Hybrid-GMS (H-GMS), a distributed scheduling algorithm for heterogeneous half-duplex (HD) and full-duplex (FD) wireless networks that combines centralized Greedy Maximal Scheduling (GMS) at the access point with distributed queue-based random access at users. H-GMS achieves throughput optimality while significantly improving delay and fairness over classical Q-CSMA, reducing average queue length by up to 79.2× under high load with optimal weight functions.

ABSTRACT

Full-duplex (FD) wireless is an attractive communication paradigm with high potential for improving network capacity and reducing delay in wireless networks. Despite significant progress on the physical layer development, the challenges associated with developing medium access control (MAC) protocols for heterogeneous networks composed of both legacy half-duplex (HD) and emerging FD devices have not been fully addressed. Therefore, we focus on the design and performance evaluation of scheduling algorithms for infrastructure-based heterogeneous HD-FD networks (composed of HD and FD users). We first show that centralized Greedy Maximal Scheduling (GMS) is throughput-optimal in heterogeneous HD-FD networks. We propose the Hybrid-GMS (H-GMS) algorithm, a distributed implementation of GMS that combines GMS and a queue-based random-access mechanism. We prove that H-GMS is throughput-optimal. Moreover, we analyze the delay performance of H-GMS by deriving lower bounds on the average queue length. We further demonstrate the benefits of upgrading HD nodes to FD nodes in terms of throughput gains for individual nodes and the whole network. Finally, we evaluate the performance of H-GMS and its variants in terms of throughput, delay, and fairness between FD and HD users via extensive simulations. We show that in heterogeneous HD-FD networks, H-GMS achieves 16-30x better delay performance and improves fairness between HD and FD users by up to 50% compared with the fully decentralized Q-CSMA algorithm.

Motivation & Objective

• To address the challenge of designing medium access control (MAC) protocols for heterogeneous networks with both legacy half-duplex (HD) and emerging full-duplex (FD) devices.
• To develop a distributed scheduling algorithm that maintains throughput optimality while improving delay and fairness in HD-FD networks.
• To evaluate the performance gains of upgrading HD nodes to FD nodes in terms of network capacity and user fairness.
• To demonstrate that combining centralized GMS with distributed random access yields superior delay performance compared to fully decentralized Q-CSMA.

Proposed method

• Propose Hybrid-GMS (H-GMS), a distributed algorithm that leverages centralized GMS at the access point (AP) for downlink queue resolution and distributed contention at users for uplink access.
• Use fluid limit analysis to prove that H-GMS is throughput-optimal, by showing it approximates GMS in fluid limits despite not being equivalent to Maximum Weight Scheduling (MWS).
• Introduce a family of weight functions f(x) for transmission probabilities, with p_l(t) = exp(f(Q_l(t))) / (1 + exp(f(Q_l(t)))), to control aggressiveness and delay performance.
• Design variants of H-GMS (e.g., H-GMS-E) with different degrees of centralization to study fairness-efficiency tradeoffs.
• Derive lower bounds on average queue length to analytically evaluate delay performance of H-GMS.
• Conduct extensive simulations comparing H-GMS with Q-CSMA under varying traffic loads, weight functions, and network configurations.

Experimental results

Research questions

• RQ1 Can a hybrid scheduling approach combining centralized GMS and distributed random access achieve throughput optimality in heterogeneous HD-FD networks?
• RQ2 How does the delay performance of H-GMS compare to classical Q-CSMA, especially under high traffic loads?
• RQ3 What is the impact of different weight functions f(x) on the delay and fairness of H-GMS?
• RQ4 How does the number of FD users affect fairness and throughput gains in HD-FD networks?
• RQ5 To what extent does H-GMS improve fairness between HD and FD users compared to Q-CSMA?

Key findings

• H-GMS achieves throughput optimality in heterogeneous HD-FD networks by approximating Greedy Maximal Scheduling (GMS) in fluid limits, even though it is not equivalent to Maximum Weight Scheduling (MWS).
• H-GMS reduces average queue length by up to 79.2× compared to Q-CSMA under high traffic intensity (ρ=0.98) with a linear weight function f(x)=x.
• With a logarithmic weight function f(x)=½log(1+x), H-GMS still improves delay by 1.2× at moderate load (ρ=0.8), while achieving up to 52.8× improvement under high load.
• H-GMS improves fairness between HD and FD users by up to 50% compared to Q-CSMA, with H-GMS-E achieving the best fairness independent of the number of FD users.
• The delay improvement of H-GMS over Q-CSMA increases with the aggressiveness of the weight function, with sublinear and linear functions (f(x)=√x, f(x)=x) yielding 10–20× better delay than logarithmic functions.
• The performance gains are most significant under high load, where H-GMS with f(x)=x reduces average queue length by 9.8× compared to Q-CSMA at ρ=0.98.