Today Nutanix announced support for HPE ProLiant server hardware and a new consumption model called “Nutanix Go”. Both announcements support Nutanix’s position that the “enterprise cloud” should be flexible, easy to consume, and with the power of the public cloud….what I like to call the “have it your way” model.
HPE ProLiant Support
The announcement of support for HPE server hardware probably doesn’t come as a surprise to many because it’s very similar in nature to the announcement of support for Cisco UCS hardware just a few months ago. While Nutanix had OEM agreements in place with both Dell and Lenovo hardware, customers wanted the flexibility to use Cisco UCS – their existing server hardware standard, and after a validation process, Nutanix offered a “meet in the channel” procurement model where a customer buys the Nutanix software from an authorized Nutanix reseller and then buys the validated server hardware from an authorized Cisco reseller. The announcement for HPE follows this same model using select HPE ProLiant server hardware (currently DL360-G9 and DL380-G9).
While it’s safe to say that there will probably be some gnashing of the teeth regarding this announcement just like there was from the Cisco UCS one (especially in light of HPE’s recent acquisition of SimpliVity), I see it as a win for everyone involved – the customer gets another choice for server hardware and the software that runs on it, channel partners have more “tools in their tool chests” to offer best in class solutions to their customers, and vendors get to move more boxes.
As mentioned earlier, Nutanix plans to support two HPE ProLiant server models initially – DL360-G9 and DL380-G9. The DL360 is a 1U server with 8 small form factor drive slots and 24 DIMM slots. The targeted workload for this server (VDI, middleware, web services) would be similar to the Nutanix branded NX3175…things that may be more CPU intensive than storage IO/capacity intensive. The DL380 is a 2U server with 12 large form factor drive slots and 24 DIMM slots. The targeted workload for this server would be similar to the Nutanix branded NX6155/8035…things that may generate larger amounts of IO or require more storage capacity.
Nutanix will offer both Acropolis Pro and Ultimate editions in conjunction with the HPE Proliant server hardware. Starter and Xpress editions will not be available at this time. However, one interesting tidbit is the fact that software entitlements are transferable across platforms, meaning that a customer could leverage Nutanix software on an existing HPE server hardware investment (assuming it met the validated criteria) and at a later date “slide” that software on over to a different HPE server model or perhaps a Cisco UCS server at the time of a server hardware refresh, if they so chose.
Support is bundled with the software license as a subscription in 1, 3, or 5 year terms. Just like the model with Nutanix running on Cisco UCS hardware, the server hardware vendor still fields hardware concerns, Nutanix will support the software side, and when in doubt, call Nutanix – if the issue is on the hardware side, concerns will be escalated through TSA Net for handoff to HPE support.
As far as availability timelines are concerned, it should be possible to get quotes for this solution at announcement (today – May 3 2017), with the ability to place orders expected for Q3 2017, and general availability targeted for Q4 2017.
Nutanix labels Nutanix Go as “On-premises Enterprise Cloud infrastructure with pay-as-you-Go billing”. In a nutshell, a customer now has the ability to “rent” a certain number of servers for a defined term, ranging from 6 months to 5 years depending on configuration and model, with pricing incentives for longer term agreements, and billing / payment occurring monthly.
While an outright purchase is probably still the most advantageous in terms of price, there are plenty of scenarios beyond price where the flexibility of quickly scaling up or down in a short time period without keeping hardware with a 3 or 5 year lifecycle on the books…having costs fall under OPEX instead of CAPEX, “de-risking” projects with uncertain futures, augmenting existing owned Nutanix clusters, etc. Customers will have the ability to mix “rented” nodes with “owned” nodes within the same cluster, enabling a sort of “on premises cloud bursting” capability.
The pricing for Nutanix Go is structured in such a way that the TCO is supposed to be significantly less than running a similar workload in AWS while mitigating some of the “use cases” that may traditionally necessitate consuming a public cloud.
Nutanix Go includes hardware, software, entitlements, and support under one SKU. It’s priced per block, per term length, and as mentioned previously, billing and payment occur monthly. Currently, there is a minimum of 12 nodes required for an agreement which in my opinion is a bit high. I’d like to see something more a long the lines of what is the required minimum for a Nutanix cluster…something like 3 or 4 nodes that might be more attractive to small and medium sized business. On the flip side, since it is Nutanix keeping the hardware on their books and allowing the customer to rent it, I can see why they’d want a certain minimum to make it worth their while. Perhaps this will change in the future.
As far as availability is concerned, Nutanix Go is initially only available to US customers, with rollout country by country for the rest of the world in the second half of 2017.
In summary, “more choices” is always a good thing, and further proof that the “power” is in software. I’m sure many customers, both potential and existing, will find these new consumption models to be a welcome addition.