EXPERIMENTAL MODELING AND PERFORMANCE EVALUATION OF SECURITY-HARDENED VMWARE VSAN CLUSTERS UNDER HIGH AVAILABILITY AND DISASTER RECOVERY SCENARIOS

Abstract

Business continuity is a critical need for any business environment today, especially one that is virtualized, with High Availability (HA) and Disaster Recovery (DR) being two critical needs. Small and Medium-sized Businesses (SMBs) don't always have the budget and resources to implement these solutions, however. VMware vSAN is a leading Hyper-Converged Infrastructure (HCI) platform that is a potentially economical option for HA and DR, but with limited empirical research on the impact of security hardening on vSAN performance and resilience. This paper introduces an experimental modeling framework to assess the performance-security trade-off of a security-hardened VMware vSAN cluster in both normal operation, host failure, and disaster recovery scenarios. A failover & recovery vSAN testbed was created using VMware Workstation Pro with failover behavior enabled, hardened with RBAC, network isolation and firewall rules, and validated using IOmeter and CrystalDiskMark, while monitored through vCenter and validated using Veeam Backup & Replication. Preliminary results are in line with published vSAN benchmarking literature and show that the cluster provides production-relevant IOPS and latency, recovers workloads from a simulated host failure within about 1 minute, keeps security hardening overhead modest (single-digit percentages), and provides for DR objectives within typical SMB continuity requirements. The results provide a fact-based benchmark for SMB decision-makers to consider when making decisions about security and performance in HCI deployments.