Performance Isolation in Multi Tenant Cloud Data Centers
Résumé
Ensuring network performance isolation in multitenant cloud data centers is critical for maintaining Service Level Agreement (SLA) compliance while supporting diverse and dynamic workloads. Traditional approaches predominantly emphasize flow-level fairness, which, although effective in certain scenarios, often fails to address the broader needs of tenant-based fairness essential for SLA-driven environments. My research introduces a novel framework leveraging control theory and AI-enhanced decision-making to achieve adaptive and intelligent network resource management. By shifting the focus to tenantbased fairness, the proposed solution ensures that resources are allocated equitably among tenants while meeting SLA requirements under fluctuating conditions. Preliminary investigations demonstrate the potential of this approach to enhance reliability, scalability, and efficiency, addressing the limitations of fairnesscentric methods focused solely on flows. This work lays the groundwork for intelligent, automated resource management in cloud data centers, advancing the capabilities of multi-tenant infrastructure to meet future demands.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |