A Predictive Fault-tolerance Framework for IoT Systems

Lead Research Organisation: Lancaster University


Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.



Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509504/1 30/09/2016 29/09/2021
1806126 Studentship EP/N509504/1 30/09/2016 30/03/2020
Description What I have discovered is that fault-tolerance solutions in literature have been very inflexible (e.g. too many assumptions about system hardware, software, infrastructure) and target very specific IoT applications (e.g. home automation, manufacturing). Consequently, I have been designing a framework that can handle faults for any application in IoT via the inference of errors in data. My last publication, "A Microservices Architecture for Reactive and Proactive Fault Tolerance in IoT Systems", proposed an architecture for efficient and scalable fault-tolerance deployment that can support my research objectives.

I have developed an indoor agriculture system as a testbed for my research, onto which my framework can be deployed. This was chosen because "smart agriculture" is an emerging area of IoT that is geared towards efficient processes (e.g. smart irrigation).
It is also an important domain due to the future concerns of increased urbanisation and a growing world population.
It provides a realistic IoT solution with plenty of fallible hardware and software to perform an effective analysis of how fault-tolerance support can handle system failures.

I have also created my own complex event processing system, BoboCEP, to provide resilient fault-tolerance support at the network edge. It actively replicates the state of partially completed complex events to enable distributed processing that can withstand hardware failure.

Since the publication of my conference paper "Providing Fault Tolerance via Complex Event Processing and Machine Learning for IoT Systems", I have been contacted by a senior software engineer from the vertical-farming company Intelligent Growth Solutions who saw my paper presentation in Bilbao, Spain. He has supplied me with a real-world vertical-farming dataset which has helped to verify that my research can be applied to solve real-world solutions.
Exploitation Route A long-term goal for my research is to develop software that will enable developers to directly implement my proposed fault-tolerance framework, so that they can easily deploy it in their own IoT environments. The software is designed to be pluggable, so that it can be deployed in any IoT system in the form of a "microservice" (i.e. a small, self-contained software package). This will enable future researchers to apply my framework and concepts for their own research.

My software, BoboCEP, is open source and fully documented, so it is available for other researchers to use for their own projects.
Sectors Agriculture, Food and Drink,Digital/Communication/Information Technologies (including Software),Electronics,Environment,Manufacturing, including Industrial Biotechology,Security and Diplomacy

Title BoboCEP 
Description BoboCEP is a complex event processing (CEP) engine designed for edge computing in Internet of Things (IoT) systems to provide inferential reasoning and decision making using stream data. It provides fault tolerance (FT) via the active replication of partially-completed complex events across multiple instances of the software using a message broker. 
Type Of Technology Software 
Year Produced 2019 
Open Source License? Yes  
Impact A short paper has been accepted that describes how and why the software was built. I am currently aiming to use this paper in collaborative efforts with other researchers. 
URL https://github.com/r3w0p/bobocep