Instead, the company wants to abstract all the stuff that makes up modern applications and make managing infrastructure idiot-proof.
More ReadingNutanix hyper-converges upwards with bells, whistles and KVMNutanix: We're not a storage company, but, er, watch out, NetAppVMware hypervisor escape via serial port? VMware hypervisor escape via serial port.VMware doubles node count for EVO:RAIL hyperconvergencewareSimpliVity opens its doors to KVM, OpenStack
We'll explain that in a moment.
The company's now talking about a platform it's calling the Xtreme Computing Platform (XCP), which has two parts.
The first is “Acropolis”, which is now the umbrella name for Nutanix's compute and storage offerings. The company's Distributed Storage Fabric is one temple on the Acropolis and does all the fun software-defined storage things that have made Nutanix a $300-million business.
Also on the Acropolis is the “App Mobility Fabric”, tools that manage VMs but do so from an application-centric point of view. The idea is that instead of managing VMs, IT folk can manage all the hardware and software dedicated to a particular application.
That's made possible, in part, by the Acropolis hypervisor, a custom cut of KVM tuned to the company's app management ambitions. The hypervisor is said to be lightweight, speedy and suited to VMs or containers.
“We made it lean and we made it secure and easy to deploy,” Binny Gill, Nutanix chief architect, told The Reg. That's because Nutanix thinks “all the intelligence is outside the hypervisor, so you only need a cheap, lean, hypervisor.”
There's your shot at VMware, in case you've been waiting for it.
So where is the intelligence? Nutanix has also announced another new release called “Prism”. And yes, it really did call it Prism, just like the NSA!
Gill told The Reg that if Acropolis is Nutanix's left brain (the logical bit) then Prism is the creative right brain. That may not be the best analogy, because Prism is a management tool that offers a “True consumer-grade user experience“ and “an end-to-end view of all workflows” so that admins can make applications do their bidding “with no need for specialized training".
All of those nice marketing words mean, Gill said, that “if I want to create 100 VMs and the system says I do not have space, I can now saw do some analysis and some VM sprawl analysis” with the machine learning routines built into Prism.
Not many clicks later, we're promised, the problem will be addressed, without a sysadmin needing to get their hands particularly dirty.
Gill said Nutanix hopes to extend this capability so that workloads tended by Nutanix software can burst into public clouds.
There's also an intention to offer users a definition of “services” that describe a “collection of runtimes” that make up an app and/or a business process.
“Managing at the container or VM level is not right,” Gill said. “We will work at the granularity of applications and services.”
Also not right, at least for all use cases, is Nutanix's current sell-with-an-appliance-or-approved-servers modus operandi. Gill said the company is looking to do end user licence agreements of Acropolis and Prism, but only for large customers.
Just when and how those licences will appear hasn't been discussed, but The Reg is at Nutanix's .NEXT conference and will try to find out. ®
Simon Sharwood travelled to Miami as a guest of Nutanix. And is very jet lagged indeed.