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

The article explores the industry trend of migrating data‑intensive analytics applications from on‑premises to cloud‑native environments, emphasizing that the unique cost model of cloud storage demands a more detailed understanding of performance optimization.

Through an empirical study of Uber’s Presto production workload, it discovers that over 50% of data accesses are smaller than 10 KB and more than 90% are under 1 MB, indicating a highly fragmented access pattern whose financial implications differ markedly from traditional data platforms.

The case study demonstrates that conventional I/O optimizations, which ignore the financial cost of storage API calls, can result in unexpectedly high expenses when deployed in cloud settings.

Based on these observations, the paper presents logical I/O‑optimization techniques and strategies specifically designed for cloud environments, aiming to help architects design efficient I/O solutions that significantly improve cost‑performance for data‑intensive applications in the cloud.