Adaptive Serverless Resource Management via Slot-Survival Prediction and Event-Driven Lifecycle Control

📅 2026-04-07
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
This work addresses the high cold-start latency and low resource utilization inherent in static resource allocation within serverless computing environments. To overcome these limitations, the authors propose a dynamic resource management mechanism that integrates event-driven orchestration with probabilistic modeling. By leveraging sliding-window aggregation and asynchronous processing, the approach constructs a slot survival prediction model, which intelligently adapts instance idle timeouts and request queuing strategies to proactively govern function instance lifecycles. Experimental evaluation across multi-cloud platforms demonstrates that the proposed scheme reduces cold-start latency by up to 51.2% and nearly doubles resource cost efficiency, significantly outperforming conventional static provisioning strategies.
📝 Abstract
Serverless computing eliminates infrastructure management overhead but introduces significant challenges regarding cold start latency and resource utilization. Traditional static resource allocation often leads to inefficiencies under variable workloads, resulting in performance degradation or excessive costs. This paper presents an adaptive engineering framework that optimizes serverless performance through event-driven architecture and probabilistic modeling. We propose a dual-strategy mechanism that dynamically adjusts idle durations and employs an intelligent request waiting strategy based on slot survival predictions. By leveraging sliding window aggregation and asynchronous processing, our system proactively manages resource lifecycles. Experimental results show that our approach reduces cold starts by up to 51.2% and improves cost-efficiency by nearly 2x compared to baseline methods in multi-cloud environments.
Problem

Research questions and friction points this paper is trying to address.

serverless computing
cold start latency
resource utilization
workload variability
cost-efficiency
Innovation

Methods, ideas, or system contributions that make the work stand out.

slot-survival prediction
event-driven lifecycle control
adaptive serverless resource management
cold start reduction
probabilistic modeling
🔎 Similar Papers
No similar papers found.
Zeyu Wang
Zeyu Wang
PhD Student, University of California, Santa Cruz
Computer VisionMachine LearningDeep Learning
C
Cuiqianhe Du
University of California, Berkeley, Berkeley, USA
R
Renyue Zhang
New York University, New York, USA
K
Kejian Tong
Independent Researcher, Mukilteo, USA
Q
Qi He
Fordham University, New York, USA
Qiyuan Tian
Qiyuan Tian
Tsinghua University, Stanford University, Massachusetts General Hospital, Harvard Medical School
MRIDiffusion MRINeuroimagingDeep Learning