01. VMware vSAN Concept

This is part of the VMware vSAN Technical Deep Dives and Tutorials series. By using the following link, you can access and explore more objectives from the VMware vSAN.
VMware vSAN [v8] – Technical Deep Dives and Tutorials

vSAN is an enterprise solution that is fully integrated with vSphere, which means that you don’t need to install additional software. When you turn on vSAN, you get a 60-day trial license to use all of the vSAN features and can perform the following functions:
• Aggregate local storage together into a single cluster-wide shared datastore
• Utilize networking for host communications
• Enhance server hardware resource utilization
• Reduce data center requirements for power and cooling

vSAN works is by taking the local storage in the ESXi host and aggregating it all together to form a single vSAN datastore. The storage devices can be SSDs, Spinning Disks, or NVMe. Once we aggregate all of the storage together, we then use the network to provide communications between the different hosts. Having a reliable network and a fast network ensures that vSAN operates optimally. Ultimately, this allows us to maximize our existing resources and our server footprint.
The administration is performed inside of the vSphere client instead of an external tool.

vSAN Cluster Types

vSAN provides four types of deployment architectures:

Single-site vSAN cluster: This is the most common type of deployment and is typically used for a single-site deployment. vSAN is able to support between 3 to 64 hosts per vSAN cluster.
Two-Node vSAN cluster: This is typically used for a small branch office or an environment that doesn’t need a large footprint. The Two-Node vSAN cluster can run a max of 25 VMs with the Remote Office Branch Office vSAN license. To fully support this deployment, a Witness is needed to ensure quorum. The witness can either be an appliance (preferred) or a physical server. If a physical server is used, an additional vSAN license is required. The witness typically runs at the main data center or a third alternative site.
Stretched vSAN Cluster: This is typically used in a metro or campus wide site, and provides geographic deployment of vSAN. This ensures if power, cooling, or network are lost at one site, the VM workload can run at another site. A max of 20 hosts are supported per data site, and vSAN supports a max of two data sites. The witness is needed at a dedicated third witness site.
Cross-Cluster Capacity Sharing: This is typically used to share siloed storage between vSAN clusters. You can mount remote datastores from one vSAN cluster to another vSAN cluster either bi-directionally or unidirectionally.

vSAN Datastore

vSAN takes all of the local storage, aggregating it all together to form a single cluster-wide datastore. With vSAN, we just have one local datastore per cluster. As we add more local storage devices (SSDs, spinning disks, or NVMe), we can scale up the storage/vSAN datastore in our environment. We can also add more host to the environment, scaling out, which adds more compute and storage resources.

The vSphere Client has a couple of default storage policies, but we can create custom storage policies to fit our business needs. Storage policies allow us to tell vSAN how we want to structure data between hosts (RAID-0/1/5/6). vSAN can also provide file shares via SMB, NFS, and iSCSI to fit business needs.

vSAN Data Storage

Our ESXi hosts can support different types of datastores (vSAN, VMFS, and NFS). At the end of the day, they all store VM data, but how they store data is different. vSAN is object-based storage vs block-based storage with VMFS, and file-based storage with NFS. An example of an object would be a VMDK. If you have a VM that has two VMDKs, you’d have two VMDK objects. We then apply a storage policy to those objects. For example, VMDK 1 could have a RAID-1 mirroring policy and VMDK 2 could have a RAID-5 striping policy.

Reference: VMware(by Broadcome) Docs