According to the  Nutanix Prism Central Guide , you can click on any chart in a widget to expand it and view more details about the data elements.

Configure VM-Host affinity for the nodes that meet the application’s GPU requirement.  This is because VM-Host affinity allows the administrator to specify which nodes a VM can run on or must not run on 1 . By creating a VM-Host affinity rule that binds the application VM to the three nodes that have the physical GPU, the administrator can ensure that the VM will always run on a node that meets its requirement, regardless of any HA or migration events. This also allows the other nodes in the cluster to host other VMs without any restrictions.

According to the Nutanix Support & Insights web search result 1 , erasure coding is a data savings technique that utilizes stripes and parity calculation in a Nutanix cluster. Erasure coding increases the usable capacity on a cluster by reducing the replication factor of data blocks. Instead of replicating data, erasure coding uses parity information to rebuild data in the event of a disk or node failure.  Erasure coding can save up to 50% of storage space compared to replication factor 2, and up to 75% compared to replication factor 3 2 .

Prism Central is a multi-cluster manager that provides a single, centralized management interface for Nutanix environments 1 . Prism Central can be deployed in different configuration scenarios, depending on the network and security requirements of the environment. Some of the valid scenarios are:

Environments use Network Address Translation (NAT): NAT is a method of mapping one IP address space to another by modifying network address information in IP datagram packet headers while they are in transit across a traffic routing device 2 .  NAT can be used to enable communication between Prism Central and Prism Elements that are in different networks or subnets 3 .  For example, Prism Central can be deployed in a public cloud and use NAT to access Prism Elements that are in a private data center 3 .

Prism Elements and Prism Central are in different subnets: A subnet is a logical subdivision of an IP network that allows multiple networks to share a single physical network 4 .  Prism Elements and Prism Central can be in different subnets as long as they can communicate with each other through routing or NAT 3 .  For example, Prism Central can be deployed in a management subnet and access Prism Elements that are in different application subnets 3 .

Environments do not have Internet access: Internet access is not required for the deployment of Prism Central, as long as the environment meets the prerequisites and considerations for installing or upgrading Prism Central. For example, Prism Central can be deployed in a dark site, which is an environment that does not have Internet access or has restricted Internet access. In this case, the administrator needs to manually enable microservices infrastructure and download the required software packages from another source.

According to the Nutanix Support & Insights web search result 1 , if the owner Controller VM becomes unavailable, the address moves to another Controller VM, ensuring that it is always available. This IP address is also used as a cluster-wide address by clients configured as part of Nutanix Files and other products. Therefore, the host will automatically redirect I/O and VMs will continue running without any interruption.