Today Nutanix is announcing its new 3.5 release of their own software.
As a partner I had the chance to attend a quick preview of this new version, and I should say the news are really interesting. First of all, it’s important to note that each new feature will be available to all customer with an active support contract. Avoiding never-ending upselling towards existing customers is a wise move, and will sure make customers happy!
There are further performance improvements, and at these levels of performance are not easy to achieve, but most of all there are some news that deserve further attention.
Nutanix Elastic Deduplication Engine
Since previous versions, Nutanix had some compression techniques (post-precessing, not in real time) to lower the amount of data saved to storage. In this new version everyone was expecting some further evolution, and deduplication was a possibility. So ,here it is:
It’s a realtime process, and can be applied to the first two level of storage, that are RAM and SSD, inside the controller VM (CVM). In this way, several benefits are achieved: first, an even better usage of storage resources, since you can save more data in the same physical space. Second, to reduce write activities on SSDs extends their life time. Being a pure software solution, this will be available on every Nutanix model regardless of the hypervisor in use, being it vSphere, KVM or the just announced Hyper-V. It’s already in the roadmap the availability of deduplication also for mechanical disks.
This is the completely new and revamped management interface, made of a new GUI and the new REST APIs. in regards of the GUI, the old one based on Flash was really nice looking, but was nonetheless dismissed and now there is a totally new HTML5 based interface.
This means it can now be used with any existing browser, and offers a better integration with the underlying management engine. In fact, since the complete command list is available both on the GUI and via the REST APIs, it’s easy to suppose the backend engine is common to them, and the choice between GUI or APIs only depends on customers needs. If you need further informations about this new feature, have a look at the presentation made by Steven Poitras during the last Tech Field Day 9:
VMware Site Recovery Manager SRA
After attending the official training, and being able to use the Nutanix shell with its commands for replica management, I thought that because the way it was built it would have been really easy to add an official SRA to support Site Recovery Manager. And in fact, here it is the Storage Replication Adapter. Even if you can manage replica and failover at the storage layer directly with Nutanix, SRM support adds orchestration capabilities for Disaster Recovery, and the official support for a solution that is much appreciated by many companies.
Nutanix was born with VMware vSphere, but since the beginning it was always said that was a choice based on the fact it was (and still is) the most common hypervisor. The Nutanix architecture however, based on commodity hardware and a virtual machine representing the heart of the solution, is per se agnostic, and can be executed on different hypervisors. After announcing KVM support some months ago, now is the turn of the third hypervisor, Microsoft Hyper-V.
The management is common to all the hypervisors, and all the existing and new features are available regardless of the chosen hypervisor. So, you can choose Nutanix whatever is the hypervisor you want or need, and also you can swap between them without changing the underlying Nutanix technology. Right now the Hyper-V support is classified as “Tech Preview”, but I’m pretty sure the complete support will be available in the upcoming months. An interesting note is the support for SMB 3.0: I have no further informations, but seems the SMB share are going to be served directly by the Controller VMs. If that will be the case (and seems logic it is this way), it will be another proof that the heart of the Nutanix solution is their native filesystem named NDFS, and the filesystems exposed to the hypervisors like NFS or SB are nothing more than a “view” on their native filesystem.