[Paper Review] A Delay Optimal MAC and Packet Scheduler for Heterogeneous M2M Uplink

Akshay Kumar, Ahmed Abdelhadi+1 more|arXiv (Cornell University)|Jun 21, 2016

IoT Networks and Protocols17 citations

TL;DR

This paper proposes a delay-optimal, proportionally fair multi-class packet scheduler for heterogeneous M2M uplink traffic at the Application Server (AS), using iterative convex optimization to time-share preemptive priority policies and maximize system utility. It further introduces a low-complexity distributed joint channel allocation and scheduling scheme that converges quickly to centralized performance with minimal overhead, outperforming WRR, max-weight, and priority schedulers in simulations.

ABSTRACT

The uplink data arriving at the Machine-to-Machine (M2M) Application Server (AS) via M2M Aggregators (MAs) is fairly heterogeneous along several dimensions such as maximum tolerable packet delay, payload size and arrival rate, thus necessitating the design of Quality-of-Service (QoS) aware packet scheduler. In this paper, we classify the M2M uplink data into multiple QoS classes and use sigmoidal function to map the delay requirements of each class onto utility functions. We propose a proportionally fair delay-optimal multiclass packet scheduler at AS that maximizes a system utility metric. We note that the average class delay under any work-conserving scheduling policy can be realized by appropriately time-sharing between all possible preemptive priority policies. Therefore the optimal scheduler is determined using an iterative process to determine the optimal time-sharing between all priority scheduling policies, such that it results in maximum system utility. The proposed scheduler can be implemented online with reduced complexity due to the iterative optimization process. We then extend this work to determine jointly optimal MA-AS channel allocation and packet scheduling scheme at the MAs and AS. We first formulate a joint optimization problem that is solved centrally at the AS and then propose a low complexity distributed optimization problem solved independently at MAs and AS. We show that the distributed optimization solution converges quickly to the centralized optimization result with minimal information exchange overhead between MAs and AS. Using Monte-Carlo simulations, we verify the optimality of the proposed scheduler and show that it outperforms other state-of-the-art packet schedulers such as weighted round robin, max-weight scheduler etc. Another desirable feature of proposed scheduler is low delay jitter for delay-sensitive traffic.

Motivation & Objective

• To address the challenge of heterogeneous M2M uplink traffic with diverse delay, size, and arrival rate requirements.
• To design a QoS-aware packet scheduler that maximizes system utility while ensuring low delay jitter for delay-sensitive traffic.
• To jointly optimize MA-AS channel allocation and packet scheduling with minimal information exchange.
• To develop a distributed optimization framework that converges quickly to centralized optimal performance.
• To validate the proposed scheduler's superiority over existing schedulers like WRR, max-weight, and priority-based schemes.

Proposed method

• Classifies M2M uplink traffic into QoS classes based on delay budget, payload size, and arrival rate.
• Uses sigmoidal utility functions to map delay requirements to service utility for each class.
• Models the optimal scheduler as a time-sharing combination of all preemptive priority policies to achieve any work-conserving average delay.
• Employs iterative convex optimization to solve for optimal time-sharing fractions, reducing computational complexity.
• Proposes a distributed algorithm where MAs and AS iteratively solve local optimization problems while fixing others' schedules.
• Relaxes integer subcarrier allocation constraints for upper-bound analysis and validates convergence to optimal solution.

Experimental results

Research questions

• RQ1 How can a delay-optimal scheduler be designed for heterogeneous M2M uplink traffic with diverse QoS requirements?
• RQ2 Can the optimal scheduler be computed efficiently by time-sharing between preemptive priority policies?
• RQ3 How can joint MA-AS channel allocation and packet scheduling be optimized with low complexity and minimal information exchange?
• RQ4 Does the proposed distributed optimization framework converge quickly to the centralized optimal solution?
• RQ5 How does the proposed scheduler compare in performance to state-of-the-art schedulers like WRR, max-weight, and priority-based schemes?

Key findings

• The proposed iterative scheduler achieves near-optimal system utility and outperforms WRR, WFS, max-weight, and priority schedulers in simulation.
• The distributed optimization framework converges quickly to the centralized solution with minimal information exchange overhead.
• Subcarrier allocation dynamically adapts to traffic load, assigning more subcarriers to MAs with high eSM traffic and low SM density, such as MA2.
• The system utility of the distributed scheme is very close to the upper bound obtained via real-valued subcarrier allocation, indicating near-optimality.
• The scheduler achieves near-minimal delay jitter for delay-sensitive traffic, while slightly increasing jitter for delay-tolerant traffic.
• The optimal scheduler is computationally feasible due to iterative decomposition of the large-scale optimization problem into smaller subproblems.