Everyone wants to understand how real applications interact with the devices they develop, but what’s the best way? This presentation will show how you can use BPF to analyze application workload interactions with your device in various systems. It will also go over examples of analysis made possible through this data collection.

Modern data centers invariably face performance challenges due to the rising volume of datasets and complexity of deep learning workloads. Sizeable research and development has taken place to understand AI/ML workloads. These workloads are computationally intensive, but also require vast amounts of data to train models and draw inferences. The impact of storage on AI/ML pipelines therefore merits additional study.

For the past 10 years, Backblaze has collected SMART statistics and failure data for over 250,000 hard drives and SSDs. Over that time, for over 100+ drive models, we have reported on drive failure rates as it relates to drive size, manufacturer, temperature, drive class, drive type and more. For this session will deliver our latest research on these attributes including any relevant trends and observations.