CLOUDADV system uses zero-shot LLMs to cut cloud VM costs by over 50%

Researchers have developed CLOUDADV, a system designed to optimize cloud virtual machine instance sizing by reducing overprovisioning. The system utilizes zero-shot foundation models for time-series forecasting and generates recommendations across various planning horizons. A case study demonstrated that CLOUDADV could significantly reduce simulated monthly costs for production VMs, achieving over 50% savings while maintaining a low exceedance rate. AI

IMPACT Optimizes cloud resource allocation, potentially reducing operational costs for businesses utilizing cloud infrastructure.

RANK_REASON This is a research paper describing a new system and its evaluation. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.AI →

• Azure
• Chronos 2 Forecasting Model
• CLOUDADV
• Gerlando Gramaglia

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →