Review of HA features in Hyper-V and vSphere
In this article, we will review the features of HA in Hyper-V and vSphere. HA stands for High Availability. In this article, we will examine the HA methods in these two products to find out which works better in preserving the VMs when the server faces problems.
In virtual environments, HA is considered an important function, and without HA, if a failure occurs in the host server, it will lead to the failure of virtual machines. HA in Microsoft Hyper-V and VMware vSphere have the same purpose, but they perform differently.
VMware vSphere has this capability through a service called HA. In VMware, a cluster has been created for this purpose, in which all ESXi hosts are located, and all related items are created and managed in vCenter.
When you create a cluster in your vSphere and do HA settings, one of the ESXi hosts in that cluster is considered as the primary host. This main host monitors the status of other hosts and is responsible for managing the connections of the hosts with vCenter. The main host manages the other ESXi hosts in the cluster to see if their network connection is established or not. If a host ever stops sending signals to the primary host, then the primary host must determine whether the outage is due to a simple connection issue or if the host is down altogether. Checking this issue is done by checking the data storage location, whether the host is still connected to its data store or not, if so, the host is still considered functional. For this reason, the main host monitors the VMs. If a VM is powered off, it powers it on another host.
One of the nice things about the way VMware handles the VM reboot process is that it allows network administrators to prioritize their virtual machines. In this way, the network administrator has the possibility to prioritize his virtual machines to go online. When high-priority virtual machines are restarted with HA settings, low-priority virtual machines are also restarted until all virtual machines are running and new host resources are left.
Microsoft’s approach to HA
In Microsoft, the HA approach is somewhat similar to VMware. However, each company has its own nuances. The major difference is that in VMware, HA capability is considered as a virtualization feature, while in Microsoft, it happens through Windows Failover Clustering. This feature in Microsoft allows you to use HA for different things, not only virtual machines. The Failover Clustering feature runs on every Hyper-V host inside the cluster. Unlike VMware’s approach which requires vCenter, Microsoft does not require any management server.
Microsoft uses a model called quorum to determine when a connection must occur. The performance of this model can be significantly different depending on which versions the network administrators are using. In general, the quorum model is designed to ensure that most of the hosts are available, if a host is called a node in a cluster as a result of a network outage, it tries to host quorum model prevent this situation.
It should be said that there are many similarities between these two virtualization products in this comparison. While there are differences, it does not prove that the uniqueness of one company is the reason for its superiority over another company
