Green or Fast? Learning to Balance Cold Starts and Idle Carbon in Serverless Computing

2026-02-27

AI Summary

This work addresses the trade-off between cold-start latency and carbon emissions from idle keep-alive instances in serverless computing, which is further complicated by time-varying grid carbon intensity and fluctuating workloads, rendering traditional static keep-alive strategies inefficient. To tackle this challenge, the paper presents the first unified framework that integrates carbon-awareness and delay sensitivity into dynamic keep-alive decisions, formulating the problem as a sequential decision-making process. The authors propose an adaptive scheduling approach based on deep reinforcement learning to jointly optimize cold-start probability, latency cost, and carbon emissions in real time. Experiments on real-world traces from Huawei Cloud demonstrate that, compared to static strategies, the proposed method reduces cold starts by 51.69% and idle keep-alive carbon emissions by 77.08%, significantly outperforming existing heuristic and single-objective approaches in balancing latency and carbon efficiency while approaching Oracle-level performance.

Technology Category

• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Planning with Language Models
• Reasoning under Uncertainty: Sequential Decision Making

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

Serverless computing simplifies cloud deployment but introduces new challenges in managing service latency and carbon emissions. Reducing cold-start latency requires retaining warm function instances, while minimizing carbon emissions favors reclaiming idle resources. This balance is further complicated by time-varying grid carbon intensity and varying workload patterns, under which static keep-alive policies are inefficient. We present LACE-RL, a latency-aware and carbon-efficient management framework that formulates serverless pod retention as a sequential decision problem. LACE-RL uses deep reinforcement learning to dynamically tune keep-alive durations, jointly modeling cold-start probability, function-specific latency costs, and real-time carbon intensity. Using the Huawei Public Cloud Trace, we show that LACE-RL reduces cold starts by 51.69% and idle keep-alive carbon emissions by 77.08% compared to Huawei's static policy, while achieving better latency-carbon trade-offs than state-of-the-art heuristic and single-objective baselines, approaching Oracle performance.

Problem

• serverless computing
• cold starts
• carbon emissions
• latency
• resource management

Innovation

• serverless computing
• carbon-aware scheduling
• cold start optimization
• deep reinforcement learning
• latency-carbon trade-off