Shaved Ice: Optimal Compute Resource Commitments for Dynamic Multi-Cloud Workloads

2025-03-13

AI Summary

Cloud service customers face demand uncertainty risk stemming from long-term compute commitments, yet existing research lacks empirical analysis of the cost–risk trade-off in multi-cloud environments. This paper addresses this gap using three years of real-world Snowflake multi-cloud workload data to quantify, for the first time, the joint impact of demand volatility, hardware generation upgrades, and software performance improvements on compute demand. We propose a novel multi-cloud commitment optimization framework integrating time-series forecasting, stochastic optimization, and empirically derived constraints. Under a ≥99.5% capacity availability guarantee, our approach reduces annual compute costs by 18–32%, outperforming both fixed-term and on-demand procurement strategies. Our core contribution is the first verifiable, demand-driven commitment decision model explicitly designed for realistic multi-cloud workloads—grounded in empirical data and validated through production-scale analysis.

Technology Category

• Reasoning under Uncertainty: Stochastic Optimization
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Machine Learning: Hardware-aware ML

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization

Abstract

Cloud providers have introduced pricing models to incentivize long-term commitments of compute capacity. These long-term commitments allow the cloud providers to get guaranteed revenue for their investments in data centers and computing infrastructure. However, these commitments expose cloud customers to demand risk if expected future demand does not materialize. While there are existing studies of theoretical techniques for optimizing performance, latency, and cost, relatively little has been reported so far on the trade-offs between cost savings and demand risk for compute commitments for large-scale cloud services. We characterize cloud compute demand based on an extensive three year study of the Snowflake Data Cloud, which includes data warehousing, data lakes, data science, data engineering, and other workloads across multiple clouds. We quantify capacity demand drivers from user workloads, hardware generational improvements, and software performance improvements. Using this data, we formulate a series of practical optimizations that maximize capacity availability and minimize costs for the cloud customer.

Problem

• Optimize compute resource commitments for multi-cloud workloads
• Balance cost savings and demand risk in cloud services
• Maximize capacity availability and minimize costs for cloud customers

Innovation

• Dynamic multi-cloud workload optimization
• Three-year Snowflake Data Cloud study
• Practical optimizations for cost and capacity