I will provide enough background and intuition in the talk, so that it is accessible to a broad CS audience.

We live in a data-driven world as we have been realizing for some time. Everything is generating data, in volumes and at high rates, from the sensors embedded in our physical spaces to the large number of machines in data centers being monitored for a variety of metrics. The question that I pose in this talk is:

Can all this data be utilized for improving the dependability of our computing systems?

Dependability is the property that a computing system continues to provide its functionality despite the introduction of faults, either accidental faults (design defects, environmental effects, etc.) or maliciously introduced faults (security attacks, external or internal). We have been addressing the dependability challenge through large-scale data analytics applied end-to-end from the small (networked embedded systems, mobile and wearable devices) [e.g., Mobisys-25, Security & Privacy-24, CVPR-24, 23, 22, TODAES-23, Eurosys-22, NeurIPS-20, Sensys-20, UsenixSec-20] to the large (edge and cloud systems, distributed machine learning clusters) [e.g., CCS-24, Security & Privacy-24, ICLR-24, Middleware-23, OSDI-22, Sigmetrics-22, NeurIPS-22, UsenixATC-21]. In this talk, I will first give a high-level view of how data analytics has been brought to bear on dependability challenges, and key insights and achievements from our technical community. Then I will provide details of four projects where we design systems from the small to the large, to meet dependability and performance requirements, using data-driven decisions.

For the detailed part, I will discuss the following projects: (I) Small: Streaming analytics on mobile and embedded devices; (II) Small: Sensor-edge-cloud continuum; (III) Small-to-Medium: Secure decentralized learning; (IV) Large: ML workflows on serverless computing.