Securing Stella: A Platform for Real-Time Data Monitoring and Code Generation

Abstract

The main objective of this thesis was to design, develop, and test a framework capable of real-time memory event monitoring and code generation in a pre-existing computer system. Conventional event monitoring strategies function after a program finishes or slow down the system. How can we create a solution that can monitor memory events on an already existing system while it is running? We adapted hardware solutions to create Remora, a framework that relies on separate hardware to facilitate code monitoring and generation with no performance impact to the host. To illustrate the viability of Remora, we implemented a dictionary search algorithm, behaviour monitoring algorithm, and a de-compressor for both code and data and adapted them all to work in real-time. In addition, these algorithms were tested on real hardware. These demonstrations show that a real-time approach to event monitoring and code generation has numerous practical applications beyond what is discussed in this thesis.