Optimizing I/O for Data‑Intensive Analytics in Cloud‑Native Environments: Insights from Uber Presto

This whitepaper explores the growing industry shift of migrating data‑intensive analytics workloads to cloud‑native platforms. It highlights that cloud storage introduces a distinct cost model, requiring a deeper understanding of performance optimization that accounts for storage API expenses.

The study analyzes Uber’s production use of Presto, showing that more than 50% of data accesses are smaller than 10 KB and over 90% are under 1 MB, indicating a highly fragmented access pattern. Such fragmentation has different cost implications in cloud environments compared to traditional on‑premises systems.

Based on this empirical research, the paper presents logical I/O‑optimization strategies tailored for cloud deployments, aiming to help architects design efficient I/O solutions that improve cost‑performance ratios for data‑intensive applications.

Key sections cover adjusting cognition and strategies according to varying cloud storage scenarios, case‑study evidence of additional costs incurred by widely used I/O‑optimization techniques during cloud migration, and a fresh perspective on system design in the cloud computing domain.

Readers will gain a foundational case‑study‑driven framework for crafting I/O strategies specifically optimized for cloud‑native, data‑heavy workloads.