NetSys 2023, International Conference on Networked Systems 20234-7 September 2023 – Potsdam, Germany
NetSys 2023 Tutorial on Age of Information
Tutorial: Age of information (AoI)
A new performance metric for measuring freshness of information
Date: 4th September, 2023
Speakers: Jorge Torres Gómez and Nikolaos Pappas
Introduction Age of information (AoI) is revisiting increasingly important in the research community as a valuable tool for assessing the freshness of the data. The concept results are particularly important for various applications concerning tracking or remote sensing, where the requirement of fresh data is mandatory. For instance, in vehicular networks, results are critical to quickly update neighbours’ vehicles’ location. Self-driving applications must be updated about the surroundings to avoid collisions. The network parameters must be properly tuned to maximize its ability to deliver information packets promptly.
This tutorial attempts to unfold the theoretical concept and conduct hands-on activities for measuring their related metrics, as the most used ones average AoI and peak age of information (PAoI). We will survey its foundational ideas and origins to comprehend the concept intuitively. The tutorial will also illustrate its applicability in various domains accounting for link scheduling, vehicular and satellite networks, and the emergent field of molecular networks. In the tutorial, attendees will also get some training with hands-on activities implementing the measurement of AoI-related metrics.
Objectives The topic of the talk will be Age of information (AoI) in computer networks. The AoI concept is today a mature area of research with direct application to investigate and enhance network performance. Initially studied in vehicular networks, this is a widespread concept today as explored in many other network domains –Satellite networks, Wi-Fi, and molecular networks are some examples. Mastering this topic has direct practice to design channel access mechanisms to maximize the freshness of the data.
With this tutorial, we aim for attendees to accomplish the following goals:
- Explain the AoI concept and related metrics.
- Interpret the AoI concept in different network domains.
- Implement AoI related metrics to measure the data freshness in point-to-point links.
Outline The tutorial consists of three main sections covering the fundamentals, applications, and AoI-metrics’ implementation:
Part I AoI fundamentals (60min)
- Introduction to AoI: Definition of the concept and metrics as average AoI and PAoI.
- llustrative metrics calculation: Average AoI calculation for queue-systems as M/M/1, D/M/1, M/D/1.
- System with multiple sources sharing the queue.
- Policies for packet management: Selection process of packets to discard in the queue.
- Rate control optimization: Strategies to generate packets and reduce the average AoI.
- Non-linear age functions: Autocorrelation functions, Mutual information-based metric, soft updates.
Part II: Domain-specific applications (45 min)
Next, we will survey the applicability of this concept in various domains along with remarked open problems in:
- Link scheduling: Optimizing the transmission scheduling concerning AoI metrics.
- Vehicular networks: AoI is used to optimize the transmission rate of beacon packets.
- Satellite Networks: AoI is used to balance resources among users fairly.
- Channel state information (CSI): AoI as a tool to balance a cost-utility trade-off.
- Molecular Communications: AoI is used to balance the impact of transmission delay and interference produced by molecules’ diffusion process.
Part III: Implementing AoI-related metrics in wireless links (60 min)
Finally, we will conduct hands-on activities implementing modules in Simulink/Matlab to measure the AoI metrics. Through this implementation, we will follow AoI’s fundamentals more comprehensively. We will interconnect sources, queues, and counters for the measurements.
- D/M/1 queue using a previous M/M/1 system implementation.
- Wireless channel models accounting for realistic transmissions in outdoor scenarios.
- Mobility models accounting for the users’ mobility.
- Average AoI and PAoI metrics.
Requirements for the participants Participants need access to Simulink/Matlab as pre-installed software to conduct the planned hands-on activities.
Materials Attendees will get access to the slides and code in Simulink/Matlab.