CodeBERT Based Software Defect Prediction for Edge-Cloud Systems

Abstract

Edge-cloud system is a crucial computing infrastructure for the innovations of modern society. In addition, the high interest in the edge-cloud system leads to various studies for testing to ensure the reliability of the system. However, like traditional software systems, the amount of resources for testing is always limited. Thus, we suggest CodeBERT Based Just-In-Time (JIT) Software Defect Prediction (SDP) model to address the limitation. This method helps practitioners prioritize the limited testing resources for the defect-prone functions in commits and improves the system’s reliability. We generate GitHub Pull-Request (GHPR) datasets on two open-source framework projects for edge-cloud system in GitHub. After that, we evaluate the performance of the proposed model on the GHPR datasets in within-project environment and cross-project environment. To the best of our knowledge, it is the first attempt to apply SDP to edge-cloud systems, and as a result of the evaluation, we can confirm the applicability of JIT SDP in edge-cloud project. In addition, we expect the proposed method would be helpful for the effective allocation of limited resources when developing edge-cloud systems. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.