Submitted: October 2019
Abstract
KNX (EN 50090) is an open standard for domestic and commercial building automation which has been adopted by international manufacturers who provide KNX certified products to guarantee interoperability and inter-working. Despite of its applicability and benefits, KNX standard has some disadvantages such as: high acquisition cost, low data rate, security and scalability which can be resolved by IP (internet protocol). Since the IP offers high bandwidth, security and scalability, it is considered a good communication layer for a smart home standard. To overcome the shortcomings of existing KNX standard and allow the inclusion of IP as communication medium, the KNX Association is now oriented towards the development of IP solutions for building automation.
The goal of this thesis is to design a system which addresses the identified challenges (i.e. redundancy, latency and consistency) imposed by the IP based large scale building automation. System consistency, redundancy and availability are the key factors driving the design of proposed system. The proposed system comprises of three main components: IoT devices (endpoints such as lights, fans etc.), client (a software tool that allows users to communicate with the IoT devices) and servers (controls the data flow between the clients and IoT devices). The clients can only communicate with the IoT devices through servers. Depending on the type of data, the server tool uses two different mechanisms for ensuring the consistency i.e. majority-based strong consistency and eventual consistency. The model checking approach has been adopted for evaluating the dependability of the system. We build a CTMC model of the system operations for strong consistency aspect and define different properties for the evaluation. After specifying the properties in the logic PCTL, we apply model checking algorithm which can automatically decide whether the system is dependable. Additionally, we conducted different test experiments to compare the proposed system with the existing solution of strong consistency (i.e. Oracle Distributed Updates), to show the advantages that proposed solution offers in improving latency, redundancy and availability.