Data leaks and suspicious behavior detection in smart home using Taint Analysis

The proliferation of modern technology, including smartphones, computers, and Internet of Things (IoT) devices, has revolutionized convenience and connectivity. However, this advancement comes with a trade-off between user comfort and the risk of data leakage and privacy breaches. This master's thesis focuses on privacy issues within smart homes. The objective is to provide a proof of concept of whether dynamic taint analysis on its own is a viable solution to identify and prevent sensitive data leaks within a smart home environment. To achieve this, we created TaintWasm a dynamic taint analysis for WebAssembly binaries. TaintWasm is based on Wasabi a dynamic analysis framework for WebAssembly. The research unveils that efficient dynamic taint analysis is challenging to implement in many aspects. The Most notable ones are real-time analysis, portability and implicit flow handling. Though some results are encouraging and dynamic taint analysis arises as a strong asset for such systems. It appears that to be a viable solution it should be combined with other techniques to overcome real-world limitations.