Flexibility on day 1 and as you scale. The need to grow can arise for a variety of reasons with any architecture. Whether performance or capacity needs evolve, or as new host are added and network or fabric considerations arise, it is critical to choose a storage platform that allows for independent scaling of resources.

Offload storage tasks to ONTAP. With typical HCI environments, the host platform is responsible for compute tasks, storage operations and any network optimization on the client side. For example, CPU overhead needs to be considered when determining the hardware requirements of the compute nodes. Often difficult to preemptively scope, this overhead is commonly accepted as 10-15% and is dependent on the I/O profile of the workloads. Additionally, it is important to consider memory consumption. Memory overhead is mandatory and shouldn’t be compromised to maintain performance. Hosts can offset this by taking advantage of RDMA-capable NICs, improving network transfer efficiency, at additional costs. Finally, with an HCI platform, storage functions such as storage efficiency, RAID and failure tolerances, and encryption are handled by the hosts.

Storage Efficiency Although NetApp was the first to deliver deduplication for production workloads, this innovation wasn't the first or last one in this area. It started with snapshots, a space-efficient data protection mechanism with no performance effect, along with FlexClone technology to instantly make read/write copies of VMs for production and backup use. NetApp went on to deliver inline capabilities, including deduplication, compression, and zero-block deduplication, to squeeze out the most storage from expensive SSDs. Most recently, ONTAP added the ability to pack smaller I/O operations and files into a disk block using compaction. The combination of these capabilities has resulted in customers seeing savings of up to 5:1 for VSI and up to 30:1 for VDI.

Enterprise-grade data availability. The protection of data is paramount for any IT organization. Planning for workload fault tolerances requires careful consideration to ensure an adequate number of nodes are available when hosts are responsible for storage operations. As the number of faults tolerated increases, as does the need for additional hosts and the amount of storage provisioned to accommodate the required VM storage capacity.

Efficient backup and recovery operations. In addition to protecting data in the event of various faults, we must plan to backup VMs and workloads as part of regular IT operations. Snapshots capture the state of a VM at a specific point in time, including the VM's disk, memory, and settings. This allows an administrator to revert the VM to a previous state if something goes wrong, such as a failed update, configuration change or falling victim to a ransomware or virus attack. The storage consumed by snapshots should be taken into account when designing a balanced solution for VMware environments.

Wholistic business continuity capabilities. Beyond standard fault tolerance, backup and recovery, an organization must plan for various scenarios such as disasters, ransomware attacks, and data center site migrations. With host-based storage, addressing these challenges typically involves relying on a range of third-party solutions to effectively mitigate disasters and ensure business continuity. Furthermore, for scenarios that are network intensive, insufficiently sizing the networking and storage devices can lead to significant performance impacts.