Our storage is designed for mission-critical applications with high availability and high-performance requirements. We have two different independent storage clusters. Primary Storage cluster and Backup storage cluster.
The primary storage cluster is our main all-flash storage cluster which is made up of Intel high-speed all-flash SSDs. All the servers in this primary storage cluster are equipped with Intel high-speed SSDs. All the virtual machines (including OS disk and DATA disk) run on this cluster.
The backup storage cluster is equipped with high-speed datacenter HDDs and is used to store the backup and snapshot of the virtual machines which are running on the primary storage cluster.
These two clusters the Primary all-flash cluster and backup storage cluster are separate clusters.
The importance of having the two separate clusters is:
If you keep virtual machines and backups on the same cluster, the long chain of backups affects the performance of the virtual machines.
In case the primary cluster goes down, you still have the backups present on the secondary cluster. These backup images can further be used to spin the virtual machines. And once the primary server is up you can migrate the virtual machines.
The data is replicated to our storage system in two different algorithms. We give our customers to choose between the Dual Replication and Triple Replication options.
If our customers choose for the Dual replication, the data is copied two times over to the primary storage cluster and if the customers choose Triple replication, the data is copied three times over to the primary storage cluster.
On top of everything, we do not use hyper-converged systems in our data centers, however, we have separate hardware storage server clusters and we have separate hardware for computing.