Managing mixed RAID workloads has improved data protection and delivers strong performanceVMware vSAN can be improved in certain areas. In cases involving our large data stores with large VMs, we experience some latency, not during normal operation, but during database backup operations. We observed latency due to buffer issues from the top-of-the-rack switches. These issues are mostly network-related because all storage data traffic travels through the network. I have recently used Nutanix, and I observed that Nutanix provides better performance than VMware vSAN due to its data locality features. VMware vSAN is now providing data locality, but we did not use that option. If VMware vSAN provides additional features in the next release, such as the VM balancing feature called DRS on the cluster that VMware previously had, it would be beneficial. With DRS, VMs can move easily from one node to another within the same cluster. Nutanix does not provide that flexibility. When placing a VM on a cluster in Nutanix, the placement uses a balancing component. After that, the VM remains on the same host. If any contention occurs on the CPU or memory side, the VM stays in place until contention happens. If issues occur, the VM migrates to another host while transferring all objects to the same host. This is how their data locality is maintained. When a VM moves to any host, it moves with all VM objects. VMware vSAN does not currently offer this option. If a VM moves to another host, it accesses the disk object through the network, which increases latency. VMware vSAN now offers an option to select data locality, but it does not function like Nutanix. This is why some latency remains. If VMware vSAN can improve this feature, it would be very helpful and VMware would regain its top position.
