Lightweight Resource Graphs for Decision Support in VM Placement Management
DOI:
https://doi.org/10.33021/sjvm5863Abstract
In this paper, The VM placement is one of the most important aspects of data center management because it directly impacts resource utilization, energy efficiency, and service reliability. Most previous research on VM placement has been dominated by algorithmic approaches, such as heuristics, optimization, and machine learning, which generally aim to generate placement decisions automatically. Although this approach shows progress from a technical standpoint, limitations in interpretability and decision control remain challenges in the context of data center operational management. Various studies have discussed energy-based VM consolidation, host load prediction, and graph-based resource modeling. However, there is still a research gap regarding how the results of such technical analyses can be integrated into decision support systems that can accommodate management policies and managerial needs. This gap highlights the need for an approach that bridges technical complexity with more structured and easily understandable decision-making processes. This research aims to propose a decision support system framework for managing VM placement in data centers. The research method used is conceptual and design-oriented, utilizing system monitoring, lightweight resource graph abstraction, analytical indicators, and policy-based decision-making flows. This research does not focus on algorithm implementation or experimental evaluation but rather on developing a decision-making structure relevant for operational management. The research results in a conceptual framework that shows how technical resource information can be transformed into interpretable decision-making inputs. The discussion emphasizes the contribution of the decision support system-based approach in complementing existing VM placement methods with a managerial perspective. This research concludes that a decision support system-oriented approach has the potential to improve transparency, policy flexibility, and control in VM placement management, as well as open up opportunities for further research at the implementation and empirical evaluation stages.







