What question did this study set out to answer?

The research aims to optimize the placement of microservices in a way that minimizes interference and latency.

March 8, 2026Open Access

Artifact Interference-Aware Cross-Application Placement: A Multi-Objective Optimization Approach for Microservice Clusters

Key Points

The research aims to optimize the placement of microservices in a way that minimizes interference and latency.
Deploys a local Kubernetes cluster with a control plane and worker nodes
Utilizes load testing with k6 to simulate traffic and monitor performance
Implements a multi-objective optimization pipeline using causal probabilities and NSGA-II
Achieves significant reduction in latency across microservices
Demonstrates effective interference management within application clusters
Quantifies the contributions of various optimization strategies

Abstract

KindCluster Artifact: Interference-Resilient Microservices Placement (End-to-End Guide) This repository provides a reproducible artifact for evaluating interference-aware microservice placement in a local Kubernetes cluster. It includes: Cluster deployment: Kind (1 control-plane + 2 workers) + Istio + 3 microservice applications Load testing: k6 spike-load to generate latency and monitoring signals Optimization pipeline: cross-application causal probabilities (PN/PS/PNS) + STIGs + Optuna (NSGA-II) Ablation and baselines: quantify contributions and competitiveness This artifact is part of the reproducibility package of the paper accepted for publication at the IEEE International Conference in Software Architecture (ICSA) 2026. Paper Title: Interference-Aware Cross-Application Placement: A Multi-Objective Optimization Approach for Microservice ClustersAuthors: Iqra Zafar, Christian Medeiros Adriano, Holger Giese

Read Full Paperexternally

Bookmark

View Full Paper

Cite This Study

Zafar et al. (Sat,) studied this question.

synapsesocial.com/papers/69ada8dfbc08abd80d5bc4dc https://doi.org/https://doi.org/10.5281/zenodo.18900711

Bookmark

View Full Paper