Over the last few months I’ve been observing a lot of changes in Nutanix market strategy. While it’s probably for the greater good, I feel forgotten as a customer who chose their hyperconverged solution for technological reasons, but also to align with my employer financial model preferences.
First, I must admit that I’m biased towards Nutanix: I’ve been following them since 2013 and when I got an opportunity, I became a customer in 2016. At this time, we had been migrating our workloads to the cloud for few years and we were struck by a reality check: cloud was becoming expensive and was heavy in the OPEX vs CAPEX balance. Reviewing multiple scenarios, including hyperconverged players, I built a business case to return to traditional datacenters using Nutanix solution. The ROI was less than a year. The project was approved and moved forward. We delivered. Successfully.
We also got other benefits from that decision: we reduced more expenses by decommissioning remaining VMWare servers and going to AHV (which is also our technological preference as it’s KVM under the hood). One of our most important success (for us, IT people): we kept all of the agility we became so fond of by using AWS and Azure for few years.
Deploying Nutanix clusters takes merely few hours of work. As soon as we are done, we are able to instantiate workloads in the same way we would do it in the cloud. No need to think about storage, fiber fabrics, balancing loads, etc. We are also able to integrate our deployment scripts and monitoring solutions with Nutanix APIs. Keeping the environment up to date is a breeze (literally few clicks), especially if we think compared to traditional infrastructure (servers firmware, SANs, controllers, all the individual hard drives firmware craziness, fabrics, etc). Nutanix is a wonderful product.
The only missing part for me at this point was on the networking side: where were my security groups or Azure NSG in that hyperconverged awesomeness? As this was greatly improving segregation and overall security posture, it was lacking in the Nutanix solution. But, as the vendors were saying, it was on the roadmap…
After many acquisitions (PernixData, Calm.io and now other), after going public in September 2016, Nutanix continues to develop and evolve. They have been aggressively developing and adding innovative features like AFS, ABS, ACS and Calm. Sooner this year, they came up with multiple exciting products:
- Flow: The so eagerly awaited solution for app-centric network migrosegmentation.
- Beam: Multi-cloud cost optimization & compliance management service.
- Era: Automated and simplified database management. This will be launched later in 2018.
- Xi: Bridging the gab between public cloud services and on-premises Nutanix deployments.
To achieve their mission of making infrastructure invisible and elevating IT to focus on applications and services powering their business, Nutanix is moving forward as a software company, not as an IT Infrastructure company. Doing so, and presumably to satisfy their shareholders since they are now under public scrutiny, Nutanix is following their counterpart models: Microsoft, Google and so many more, in the “as a service” path. Most (if not all), of Nutanix recent products (Calm, Flow and others) will be delivered only in a subscription – pay-as-you-go – model.
While it totally makes sense for some services like Beam and Xi, I must admit that I would have preferred other options for Calm and Flow. It’s also difficult to forget vendors speeches not that far ago telling that Nutanix is always evolving their solution and that by buying the “Ultimate” package, I would get all the bells & whistles that will be available in the future. They were even adding that Nutanix was not following its counterparts (read VMWare) by selling their product under many editions with various features and multiple options. It’s clear that things changed in San Jose…
Not only do we have to pay extra money to extend our Nutanix capabilities with Flow and Calm, but we need to do so in OPEX expenses instead of CAPEX expenditure, which is even more disturbing for me. The CAPEX model was one of the main drivers in my business case to move from the cloud to Nutanix. In other words, at my actual company, by delivering new products under an OPEX model, Calm and Flow are not an option.
Don’t take me wrong, it’s great that Nutanix is extending their solutions to multiple hardware vendors and to subscription models: it will allow them to reach many potential clients who could have been blocked by the required CAPEX investment. My concern with this is what will happen to people like me, who work for companies where performance is evaluated based on EBITDA and who decided to leverage Nutanix in-house solution to limit OPEX expenses? Will they keep their CAPEX based model? For how long? What will be the limitations of these perpetual licenses and deployments?
While Nutanix is one of my favorite platforms, I’m having a hard time to figure out ways I could leverage a Nutanix deployment on a subscription basis. One of the few possibilities I may think to justify this is for the companies, who prefers OPEX and who have restrictions and cannot host their data in a public cloud provider datacenter. Going with a Nutanix subscription would allow them to have their in-house cloud. Otherwise, in a company that prefers OPEX and without data location constraints, I would probably architect around a public cloud solution, leaving behind any hardware and datacenter management while probably spending less money.
Overall, this move of Nutanix is really interesting as they are entering an interesting market. There is at least one major player in the same market: Microsoft. Azure Stack solution might be younger as an on-premises cloud solution and it might have a lot of limitations when compared to its public cloud counterpart. However, Azure platform is rich in features and Microsoft roadmap is aggressive, we know that they can deliver. I would literally love to evaluate Azure Stack against Nutanix to see how they compare in terms of features, scalability, stability and performance.
For now, I will evaluate alternatives to Flow like Illumio (maybe cheaper as license seems more flexible and less vendor lock-in) and I will probably continue to use Ansible as a configuration management solution instead of jumping in the Calm bandwagon.