Author: Membrey, Peter Loi
Title: Evaluating timing fundamentals for IoT devices
Advisors: You, Jane (COMP)
Chang, Rocky (COMP)
Degree: Eng.D.
Year: 2021
Subject: Internet of things
Hong Kong Polytechnic University -- Dissertations
Department: Faculty of Engineering
Department of Computing
Pages: xxv, 171 pages : color illustrations
Language: English
Abstract: IoT (Internet of Things) devices are changing how research is conducted and how engineers build and test new systems. The high performance to cost ratio makes these platforms extremely interesting and has lead to widespread adoption by many organisations (such as CAIDA and RIPE). However, the reason these devices are low cost is because the components are generally chosen to be 'good enough'. This means that although for general use the device is perfectly capable of performing well, none of the components are likely to be of exceptional quality. An area of particular concern is the accuracy of their clocks. Signifcant amounts of data have already been collected by IoT devices and research based on this data has been published in the literature. Much of this data contains or is based on some form of timing data such as when an event occurred or how long an event lasted. Although a signifcant proportion of this research could not have been conducted without employing IoT devices (again due to cost and logistics restrictions), the issue remains, can the timing data collected be trusted? Engineers are restricted by the requirements of their professional bodies (such as the Engineering Council of the UK) that any system they build be fit for purpose. This is diffcult to do with IoT devices as they do not come with tolerances and constraints (which is a standard practice for discrete components). This makes it all but impossible to develop solutions with IoT devices as an engineer has no way to validate whether constraints are being met. How can engineers take advantage of these platforms for capturing timing data when constraint information is not available to them? The solution to both of these challenges is for researchers and engineers to perform the characterisation and validation of the timing hardware themselves. A methodology to do precisely that, is described in the literature and boasts high levels of accuracy. However it requires advanced hardware (such as an atomic clock) and specialist knowledge that is unlikely to be available to the people who would want to take advantage of it. This thesis describes in detail a new methodology called Interrupt Driven Kernel Sampling (IDKS) that greatly simplifes the characterisation of the counter in a device (IoT or otherwise). IDKS injects a highly stable time reference into the device and leverages kernel modifcations to capture counter values in the interrupt handler. These are then passed off to an application running in user space that records the data to disk for later stability analysis. IDKS was used to characterise the counters in each of the current standard models of the Raspberry Pi computer platform (the Pi) as well as a Minnowboard Turbot. Using IDKS, it was possible to show that the counter on the Raspberry Pi offers surprisingly good stability and that all four generations of the Pi are acceptable for sub-ms levels of time-stamping. The Minnowboard's counter offered less stability than any of the Raspberry Pi's but was still acceptable for sub-ms time-stamping. It also offered insight into a number of issues in relying on the Linux PPS (Pulse Per Second) subsystem (such as delayed and missing PPS updates) that projects depending on accurate triggering need to be aware of. IDKS still requires a stable reference to inject into the device. This work addresses this by characterising the PPS signal from an IoT GPS device based on the uBlox M8Q chipset and demonstrates that its stability profle rivals that of an atomic reference.
Rights: All rights reserved
Access: restricted access

Files in This Item:
File Description SizeFormat 
5448.pdfFor All Users (off-campus access for PolyU Staff & Students only)7.98 MBAdobe PDFView/Open

Copyright Undertaking

As a bona fide Library user, I declare that:

  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.

By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

Show full item record

Please use this identifier to cite or link to this item: