Establishing Trust, Efficiency and Simplicity in Data Systems

Data systems have been at the heart of the digital revolution of the last two decades, enabling rapid advancements in nearly every domain from scientific discovery to industrial automation, healthcare, and beyond. Yet, as their importance grows, so do the challenges in ensuring trust, efficiency, and simplicity in a continuously evolving world. In this docent lecture, we will cover several key research outcomes from Data Systems Lab together with research partners. We will first present industry-defining methods that have ensured trust in cloud, stream, and serverless systems without compromising performance or simplicity. This includes work behind the Apache Flink, Portals, and OmniPaxos systems. Furthermore, we will touch upon the integral role of AI in the modern computing stack, examining current trends, challenges, and early successes from ongoing work in Orb DB, Aqua, uWheel, and UniCache. Finally, the seminar will conclude with future prospects in new critical areas in data computing and emerging opportunities when it comes to hardware acceleration.