Academic Work

PhD Electrical and Electronics Engineering

Topic: Physical Unclonability Framework for the Internet of Things

As the integration of Internet of Things in every aspect of every day life continues to increase, there is a tendency to create unified architectures with a great number of edge nodes and inherent security risks due to centralised data aggregation. At the same time, security and privacy defenders advocate for completely decentralised solutions splitting the control and the responsibility among the entirety of the network nodes. Naturally, dividing the responsibility among a great number of parties also means leaking information to a great number of parties, merely to enable their interaction.

The solution to achieving the best of both worlds is the primitive of unclonability. Unclonability is the basis of any relationship, be it human or between devices, as it provides proofs that the communicating players are indeed unique and can only exist in a single place at a time. This uniqueness also has a direct effect to the value of the unclonable asset since, evidently, no other copies exist to share this value. In the perspective of IoT, unclonability can offer strong security guarantees, distinction among otherwise identical edge nodes, and higher levels of control over the system by its owners. Unclonability has been realised on a physical level via the use of Physical Unclonable Functions (PUFs) but methods to expand it to fully formed security frameworks have not been produced.

In this work, we are attempting to set the basis for the development of a stack of unclonability protocols and methods, to enable the propagation of the unclonability primitive from the unclonable chips of PUFs, to devices, network links and eventually through to unclonable systems.

MSc Wireless Embedded Systems

Dissertation: Physical Unclonable Functions in Cryptoblocks

Physical Unclonable Functions(PUFs) are a novel concept of generating unique bit strings based on physical properties of electronic devices. By unquestionably identifying not only a specific device family but also a certain instance of that device, it is possible to create methods of verifying the source of data while protecting the privacy of the party that created it. The relevance of such methods is rapidly increasing due to the growth of ubiquitous networked devices. This work describes a signature protocol incorporating the potential of the PUFs to sign pieces of data and verify their authenticity when required. Through the demonstration of an operational prototype system, the challenges and capabilities of such architectures were explored and discussed.

MENg in Electrical & Computer Engineering

Dissertation: Priority Based Congestion Avoidance Technique (PB-CAT)

Wireless Sensor Networks (WSN) are typically consisted of hundreds of nodes which generate a high volume of network traffic. Due to the multi-hop communication architecture often used in such networks, bottlenecks are bound to appear in parts of the system. These bottlenecks, also known as network congestion, can have a grave effect on network performance and, given the constricted resources of WSNs nodes, result in the system failing to serve its purpose. Special care should be taken to avoid or alleviate the problem of congestion via specialised algorithms and methods. Our study aimed to lessen the effect of congestion in WSNs by studying, implementing and evaluating techniques of congestion avoidance.

In the context of this study, we designed and implemented an innovative congestion avoidance method which can successfully handle three different packet priorities, under the name Priority Based Congestion Avoidance Technique (PB-CAT). PB-CAT uses a mechanism of packet delay and merging in order to reduce the total transmissions and increase the optimal use of the communication medium, resulting in improved network performance.

In order to evaluate our method, we implemented it using the Contiki operating system and performed extensive simulations in the Cooja network simulator. The simulation results presented a clear improvement of the network metrics, particularly in sensor networks with a high volume of traffic, where we observed a considerable amount of enhancement in the quality of the network operation.