Nutanix All-Flash

Nutanix All-Flash only platforms have been available for a little while. With Nutanix being the leader in hyper-converged infrastructure, the all-flash Nutanix hardware makes much more sense.

Nutanix Flash only platform has been around for many years, but it was only in the form of a single version of their hardware line up.

With the latest Intel CPU in place, this has allowed Nutanix to enable All-Flash, within their entire range, from the 1000 series, up to the 8000.

With the largest SSD of 3.84TB available within all of the range, this means, even when using their smallest NX-1065S nodes, configure in a single cluster of 4 nodes and providing N+1 fault tolerance, you can get almost 20TiB of storage in 2U.

The new all-flash approach means that sizing an environment takes much less effort than before. With some agreed assumption with the customer, like how many vCPU can share a physical core and fast estimation can number of nodes can be configured.

One disadvantage for existing clients, you can not mix hybrid and all-flash nodes in a single cluster.

An advantage of the Nutanix All-Flash nodes is with an accurate sizing of virtual CPU required and their ratios; disk spaced used, the number of physical Nutanix All-Flash nodes can be very competitive against other All-Flash arrays, or indeed against their hybrid configuration.

My instinct says that the hybrid nodes will be phased out for new purchases within the next 18 months or so.

Nutanix is always looking at the next two or three generations of solid storage.

With newer solid state storage technologies like NVMe and V-NAND, the potential storage IOPS will be staggering.

The Nutanix all-flash nodes are excellent systems for anyone needing to replace internal infrastructure and wanting a fresh approach. Nutanix should be your next port of call if you want to streamline your internal systems and remove the cumbersome management of systems that require dedicated fiber switches between hosts and storage.

Not only is management easier, from one click updates and upgrades, but installation and space savings are also major advantages.

Using Nutanix own hypervisor, Acropolis can provide big savings when not having to license or renew support contracts with popular hypervisor providers.

Moreover, with integration increasing on a daily basis with Azure, moving work loads between public and private clouds will become easier than ever before.

Nutanix is starting to provide AWSAzure like features within your own data centre.

Group those three metrics into server types:

• SQL or other high intensive servers
• Medium Servers
• Small Servers

With those metrics grouped, MirrorSphere can provide a very quick estimation on some nodes and node type required.

If you want to get a good idea on what kind of Nutanix All-Flash system you may need, you will need the following information:

• Total Number of vCPUs
• Total Memory
• Total space provisioned

Flash makes things easier, but we have work on some general rules and assumptions, these normally work fine for estimating purposes.

We base our calculation on a ratio of virtual CPUs to Physical Cores. With the Nutanix All-Flash nodes, we no longer need to work out working sets for each and every server virtual server.

It is also a best practice that we provide the total capacity that you have provisioned and we do not work on any potential deduplication ratio as can be very misleading and is dependent on the application readwrite profiles and the applications own best practice.