Code for Paper U2CMigration

This dataset contains the source code artifacts accompanying the paper "U2CMigration: User-Unaware Container Migration with Predictive Analysis of Memory Dirty Pages". U2CMigration is an advanced User-Unaware Container live Migration strategy specifically engineered for efficient management of containerized workloads. Traditional live migration often introduces significant performance degradation, especially during the memory dirtying phase, or necessitates application modifications to become migration-aware. The core innovation of U2CMigration addresses these limitations by employing a machine learning-based model to predict future dirty memory pages. By leveraging this predictive analysis, U2CMigration significantly optimizes the container image transfer process, successfully minimizing both downtime and overall migration time without requiring any changes to the running application, thus achieving its user-unaware capability. This strategy is designed to drastically improve the performance and efficiency of migrating containerized applications, particularly those characterized by high memory write intensity, making it an essential component for dynamic resource management within modern cloud and edge computing environments. This repository includes all necessary scripts, modules, and configurations required to replicate the experimental results and deploy the U2CMigration framework as discussed in the original publication.