Nutanix has come a long way. Now there’s a software platform that you can roll out anywhere, modify at any time and move to where you need it. Or where it works the best for you, because of performance or simply in terms of cost. On-premises, in the cloud, a combination (hybrid cloud) or multiple clouds (multi-cloud).
In this article, we take a look at what Nutanix has become in the year 2020. Nutanix has changed a lot over the years; that is why it is a good idea to name the time, June 2020. Nutanix may transform even further in the future, and we will do our best to keep this article up to date as possible.
Hyper-Converged Infrastructure (HCI)
Nutanix is a so-called hyper-converged Infrastructure (HCI) solution. A software platform (cluster) that runs on top of different kinds of individual servers (nodes). All these servers are linked together through this HCI software platform. All processors, internal memory, hard disks (storage) and network interfaces, are bundled in one cluster on which you can run virtual machines. The powerful thing about an HCI platform is the way all applications and workloads are distributed across the hardware, to optimise performance as much as possible. There is also redundancy built-in by dividing data and workloads across multiple servers. As soon as one of the nodes fails, the availability of the platform and applications is not affected.
A cluster can be expanded with new nodes. It is easy to add or remove servers from the cluster. This allows the Nutanix infrastructure to grow with the company. In the end, this is not unique, because this is the basis of a hyper-converged infrastructure. Nutanix performs well as a hyper-converged infrastructure for years now, in the more recent years, it has been more about broadening the platform.
Nutanix is now a pure software company
Nutanix is now a pure software company, but this hasn’t always been the case. It once started with its own hardware appliance on which the Nutanix software ran. By buying multiple Nutanix appliances (nodes), you could build your own hyper-converged infrastructure. The company is now so big that they no longer need to develop and deliver the hardware themselves. The major hardware manufacturers are now lining up to partner with them. Cisco, Dell, Fujitsu, HPE, Huawei, Lenovo and a few others all offer hardware certified by Nutanix. It is also possible to combine hardware from different suppliers to build one large cluster. This is particularly interesting for large companies. They often have a contract with a hardware supplier with a fixed discount and SLA agreements. They can now also get Nutanix hardware within their existing contract.
At the end of the day, everybody needs to do what they do best. Nutanix is great in developing an HCI software platform, but not in building the best hardware. When companies choose for Nutanix, they can call Nutanix for software issues, but call their hardware supplier for hardware related incidents. By doing it this way, customers have the best experience and the best support.
Is Nutanix making the public cloud private or the private cloud public?
If we dive into the Nutanix platform, we see that the cluster is equipped with a traditional hypervisor. This hypervisor is developed by Nutanix and bears the name Acropolis Hypervisor, often referred to as AHV. The cluster is centrally controlled from one single pane of glass, called Nutanix Prism. Prism gives the cluster the characteristics of a public cloud because everything can be configured and controlled from one console or through APIs.
Until just over a year ago, Nutanix was primarily an on-premise solution, with the characteristics of the cloud. When we spoke to Nutanix founder Dheeraj Pandey in 2019, he already said that support for the public cloud wouldn’t take long. Today the lines between the public cloud, private cloud and on-premises environments have more or less disappeared.
You can now build a Nutanix cluster wherever you want. It can be on-premise, in a Nutanix datacenter or in the public cloud of Amazon Web Services (AWS), Microsoft Azure or Google Cloud. Or in a combination of datacenters and public cloud vendors. If you login to Nutanix Prism there is no difference, the cluster is just getting bigger and bigger, and it doesn’t matter where the environment is running. You can mix public cloud, private cloud and on-premises with Nutanix. It doesn’t matter; it just works, it’s a hybrid cloud.
The cloud has different flavours
For customers who want to go to the cloud with Nutanix, there are, as mentioned earlier, several possibilities, but within those possibilities, there are also different flavours. A customer can choose the Nutanix cloud. In Europe, Nutanix has datacenters in Frankfurt and London, from where it offers Nutanix environments. At the press of a button, you can expand an existing on-premise environment or replicate it to the Nutanix cloud to protect against a disaster.
The same applies to Amazon Web Services. Nutanix has certified the hardware of AWS so it can run directly on the bare metal servers of AWS with its own Acropolis hypervisor. As a result, it creates flexibility and the cluster can easily be extended to AWS. You can expand your cluster to AWS with one click of a button. This can be to a multi-tenant Nutanix environment managed by Nutanix, or your company’s Nutanix environment that is running in your company’s AWS customer account.
Nutanix in the public cloud on other hypervisors
If you are already a customer of AWS, Azure or Google Cloud, you can link that environment in Prism. This makes it possible to extend your Nutanix environment to the public cloud. As just mentioned, this works very easy in AWS, because the hardware of AWS is certified by Nutanix.
For Google Cloud and Microsoft Azure, it is, unfortunately, a bit more complicated. On these platforms, Nutanix has less freedom and has to work with the hypervisors of Google and Microsoft. It is still possible to set up a Nutanix environment in those clouds, but they run without the Nutanix Acropolis Hypervisor. This has the disadvantage that when you want to migrate a VM from on premises to, for example, Microsoft Azure, you first have to convert the VM to a Hyper-V format. This is supported out-of-the-box by Nutanix and not a real problem. It only takes more time. So for disaster recovery purposes where you need to move and retrieve a VM as quickly as possible, it can be an issue. If you want to be able to switch immediately, then AWS is more suitable at this moment. Maybe Nutanix will strike a partnership with Azure or Google in the future so that they can use the Acropolis hypervisor, but that is not the case at this moment.
Is it smart to run everything in the cloud?
Many mid-market companies who are rapidly replacing their legacy applications will think: why not just put it all in the cloud? Of course, we checked with Nutanix to see if that’s a wise choice. They told us it is possible to run Nutanix entirely in the cloud; they have many smaller companies that skipped the on-premises step and went straight into the cloud. However, experience shows that they will come back to this decision after a couple of months.
In almost all cases, it is cheaper to choose a hybrid cloud strategy. The vision of Nutanix is also that we go to a hybrid cloud world. Having your datacenter or renting datacenter capacity with your own hardware is simply cheaper. Especially now that all hardware manufacturers have also developed pay as go models, based on the cloud model, but for their physical servers.
In some industries, however, you have to deal with peak loads. An extensive administration system needs more compute at the end of the month. Retail is busier around the holidays than in January and February, and there are many more examples. The cloud is very suitable for such scenarios. Peak loads can easily be handled with a hybrid cloud model. You can run the baseload on-premises and the peak loads in the public cloud.
The public cloud is also very suitable for disaster recovery. You can continuously synchronise workloads to the cloud, where they are stored but not activated. You then effectively only pay for the data storage. If a failure occurs, the VMs can be switched on quickly, and operations can resume, and only then it will cost money.
Of course, the cloud can also add value, because you need to deliver a particular service to the other side of the world or the service is active worldwide and has to be quickly accessible everywhere. You can run a part of the business in the cloud and a piece on-premises.
The questions you run into pretty quickly is: What do I put on-premises? What do I put in the cloud, but above all, which cloud and under which conditions?
Nutanix Beam advises you which workloads to run where
In addition to the infrastructure, Nutanix has developed several SaaS solutions to make customers’ lives easier. Beam is such a service. With Beam, you can make your entire infrastructure transparent from a cost perspective. Beam analyses all workloads in your cluster: how much compute, memory and data traffic do they use and how do the data streams flow. It also looks at what your infrastructure looks like, whether you’re running on-premises or in the cloud.
Based on this, it can advise very precisely whether your infrastructure is set up correctly, or whether it is smarter to move a particular workload on-premises or perhaps to another cloud provider because of cost, performance or security and compliance aspects. Beam knows the cost of all cloud components. Especially the egress piece can be the decisive factor. Egress costs are incurred as soon as you move data out of the cloud onto the Internet. You pay egress costs per gigabyte. If you transfer a lot of data out of the cloud, the costs can increase rapidly. That’s what you want to avoid at all times.
If your company is going to roll out a new application or service, it is, therefore, wise to first make a good business case, taking all factors into account, before you choose to roll it out without any consideration in a cloud environment. Then it is often better to run it on-premises first so that Beam can analyse it and only then decide where exactly to put it. On-premises is almost always the cheapest option.
In addition to advising on costs and workloads, you can tell Beam that your company has specific security or compliance requirements, or that you have precise infrastructure specifications that you need to meet. These can then be included in the advice as well.
Finally, the behaviour or consumption of applications can change. Infrastructure requirements may also become stricter, and cloud providers may raise or lower their prices. Beam will respond to all these situations. So it is possible that after three months you suddenly get different advice than you get today.
At first glance, the price tag of hyper-converged infrastructure is not very attractive. But this applies to all HCI providers in this segment. What is essential in a business case and an analysis of whether an HCI is suitable for your company’s project or infrastructure is that you take all components into account. Too often, people compare apples with oranges in this case. If you compare an HCI solution to the cost of purchasing a server farm, an HCI is always much more expensive. Ultimately, the physical hardware is only 30 percent of the cost of running a datacenter. The other expenses are in the operational part, the amount of rack space, cooling, power and finally people. The people that have to manage the server farm are usually not cheap, and by automating a lot of management, costs come down quickly.
Nutanix also takes care of a large part with its HCI. The management is done through a single pane of glass. For managing the infrastructure, many things are automated or simplified to a simple push of a button. As a result, much less time is required to maintain the infrastructure. Also, with an HCI, you need less hardware than if you were running everything on separate servers. The costs for cooling, power and rack space will decrease as a result. Comparing pure iron is, therefore, not wise.
In many European countries, there is a large SME and SMB market, in those countries more savings can be achieved through managed service providers and shared service centres.
Multi-tenancy for managed service providers and shared service centers
For the average enterprise organisation, it is not a problem to purchase and set up a Nutanix infrastructure. In the end, the costs are less restrictive for them, because their projects are much larger than for SMBs or non-profit organisations. Nutanix, like many other suppliers, has an extensive network with partners and managed service providers that offer Nutanix as a reseller.
Nutanix is a multi-tenant product. Multiple customers can run their workloads and applications on a single infrastructure but still be completely separated from each other. The multi-tenancy is build into the core of the product. The internal networks are also entirely separated by micro-segmentation so that a customer can never have (internal) access to other customers data. This makes it possible for managed service providers to set up large Nutanix clusters in their datacenters and resell them to small or medium businesses or enterprises. Even then, it is still possible to expand a cluster to the public cloud. Ultimately, these managed service providers do the same as Nutanix does with its datacenters, but on a smaller scale.
Nutanix also sees that many organisations with shared service centers benefit from the multi-tenant support. For example, some municipalities collaborate and run the applications and workloads of several cities or healthcare institutions on a single infrastructure. Thanks to the multi-tenant support, the applications and data remain strictly separated, but they can achieve benefits through joint procurement.
Nutanix is still a relatively new player for many organisations, but it one of the few companies to come up with a good working and affordable hybrid and multi-cloud solution. A couple of years ago many people thought everything would become cloud-only, now it is clear that from a cost perspective this is not feasible. Hybrid cloud seems to have the future for the time being, and the number of players active in it can be counted on one hand. Also, Nutanix has a working product, and a rich customer portfolio, where others are still mostly talking about hybrid and multi-cloud and bringing in customers has yet to begin.
Nutanix is not there yet, and it will certainly have to prove itself and compete against players like Google, Red Hat, VMware and many others. It’s not easy, but it’s not impossible either. The company has certainly proved this in recent years with more than 16,000 customers worldwide.