Two-sided Delay Constrained Scheduling: Managing Fresh and Stale Data

TL;DR

This paper introduces a novel two-sided deadline framework for wireless packet scheduling, optimizing energy and completion time while managing both freshness and staleness constraints, relevant for security and real-time applications.

Contribution

It formulates and solves the optimal offline scheduling problem with two-sided deadlines, extending traditional models to include post-transmission constraints.

Findings

Optimal offline scheduling solutions under two-sided deadlines.

Duality between rate and energy in scheduling.

Framework applicable to security-critical and real-time systems.

Abstract

Energy or time-efficient scheduling is of particular interest in wireless communications, with applications in sensor network design, cellular communications, and more. In many cases, wireless packets to be transmitted have deadlines that upper bound the times before their transmissions, to avoid staleness of transmitted data. In this paper, motivated by emerging applications in security-critical communications, age of information, and molecular communications, we expand the wireless packet scheduling framework to scenarios which involve strict limits on the time after transmission, in addition to the conventional pre-transmission delay constraints. As a result, we introduce the scheduling problem under two-sided individual deadlines, which captures systems wherein transmitting too late (stale) and too early (fresh) are both undesired. Subject to said two-sided deadlines, we provably solve the optimal (energy-minimizing) offline packet scheduling problem. Leveraging this result and the inherent duality between rate and energy, we propose and solve the completion-time-optimal offline packet scheduling problem under the introduced two-sided framework. Overall, the developed theoretical framework can be utilized in applications wherein packets have finite lifetimes both before and after their transmission (e.g., security-critical applications), or applications with joint strict constraints on packet delay and information freshness.