ICPS: Real-Time Resource Configuration for Cloud Serverless Functions Considering Affinity

TL;DR

This paper tackles the challenge of minimizing serverless workflow response time in the presence of branching structures and inter-function latency. It introduces ICPS, which combines LSTM-based concurrency prediction with pre-warming and affinity-aware placement to reduce cold starts and network delays. The approach is decomposed into three prediction strategies, three deployment strategies, and three routing strategies, all orchestrated by a four-module system, and validated against multiple baselines using a real dataset and platform simulation. Results show that ICPS consistently improves both response efficiency and resource utilization, particularly under high network delay, highlighting its practical impact for scalable serverless workflow optimization.

Abstract

Serverless computing, with its operational simplicity and on-demand scalability, has become a preferred paradigm for deploying workflow applications. However, resource allocation for workflows, particularly those with branching structures, is complicated by cold starts and network delays between dependent functions, significantly degrading execution efficiency and response times. In this paper, we propose the Invocation Concurrency Prediction-Based Scaling (ICPS) algorithm to address these challenges. ICPS employs Long Short-Term Memory (LSTM) networks to predict function concurrency, dynamically pre-warming function instances, and an affinity-based deployment strategy to co-locate dependent functions on the same worker node, minimizing network latency. The experimental results demonstrate that ICPS consistently outperforms existing approaches in diverse scenarios. The results confirm ICPS as a robust and scalable solution for optimizing serverless workflow execution.

Figures (9)

• Figure 1: An example of workflow application
• Figure 2: System architecture
• Figure 3: Structure of LSTM network
• Figure 4: FPCG strategy
• Figure 5: BPCG strategy