What is our primary use case?
One goal was to automate things. We had a lot of tools, but we needed a centralized tool. Calm helps us to centralize the deployments of our VMs.
We have a subsystem installed on Nutanix and we have blueprints for setting up this subsystem very easily. Also, for Kubernetes clusters, we use now CaaS from SUSE and we also create Kubernetes clusters with Calm. Our strategy is to make blueprints for all the virtual machines environments. It's an ongoing process.
How has it helped my organization?
Our first project was to create subsystems. This was really an accelerator because we have three environments and over 50 machines. Once we had a sub-template, it was very easy to migrate to Nutanix, to set up a system. Before Nutanix it took days and now it's maybe one or two hours. It's really fast when you use these templates. It creates all the preconditions for an installation. And with that, we were really able to move the system very quickly to this new platform.
The solution automates application management to a single platform, but we're still working on it.
Our goal is the standardization which Calm makes possible. It's important, from a strategic point of view. We would ultimately like to achieve "infrastructure-as-code" so that we can always create an environment as it initially was. It would be like Kubernetes or container-based where you can destroy something and build it again and it's like it was before. When you have a platform where you can automatically create things, you are sure that nobody will manually change something in it. It's all managed with this framework, and you are sure that when when you need to create the same system it will work, because it is all scripted. The whole "cookbook" for making that machine is there. This is also a requirement: that nobody goes on a virtual machine and installs something manually. It must be scripted with Calm. That gives you insurance that you can build the same system again. For us, that's really the future: infrastructure-as-code.
This is also a good way for creating the same machine on the cloud, or wherever you want, and to be assured it will run because the building of the machine is in the script.
Also, the solution’s support for scripts, API, and domain specific language has reduced the IT man-hours to deploy and support applications. It's hard to estimate how much time it has saved us, but I would say around 60 percent. We are new on the Nutanix platform and we have not created a lot of the blueprints ourselves. Another company helped us to accelerate that. We went into production with it last year and we see the capabilities that Calm gives us.
Before Calm, we didn't have a specific tool for orchestration. We had some templating things, but they were spread out over various technologies. Now, we have one, centralized solution to manage all the VMs that we have. This is the strength of Nutanix, that you have one starting point where you can do everything. You have all the tools in one platform. Before, we had one tool for this process and another tool for that process. It's helping us a lot.
Calm has also enabled us to react faster to the changing needs of our business. That brings me back to the subsystem I mentioned earlier. We were thinking we would need more time to migrate it, or that we might need to create a sandbox system for testing. But with the subsystem, it was very quick. Calm helped us a lot to make it happen.
Also, when it comes to cluster systems, we work with the open source version of Couchbase. It's very easy to create a Couchbase cluster. Similarly with Jenkins, we have blueprints for DevOps. If they need a Jenkins environment, we can easily scale out for our Jenkins workers. It really makes life easier because we have a GUI and can scale out. We can say, "Okay, we need two more slaves," and it happens. It really accelerates things.
What is most valuable?
The scripting, where you can use libraries, is a valuable feature. We don't really make the blueprints, as we have a third-party company that makes them for us. But it enables calling APIs in the blueprints. When we create a machine, we use IPAM from Infoblox and we can get an IP address. It's one platform to script and we can then use all the APIs to complete the scripts. It gives us a central management tool from which we can do a lot of things automatically.
Also, it's easy to use, overall. I'm a Linux guy, so a lot of it is familiar to me. I feel comfortable when I use it. It's not really hard or complex.
And when you have applications that can run on more than one machine, you can easily use blueprints to scale out the infrastructure. You can start with two web front-ends, a web service and then you say, "Okay, I need a third one and a fourth one." This is very easy. It's one click and you can scale it, but you must also script it. It only gives you the framework to do that. So for performance, you can use Calm to scale out and scale in.
But the Nutanix platform also helps you find out if you have some performance problems or oversized machines. But to resize it, it's more that you would use playbooks in Nutanix for that, and not Calm.
It's also a very good tool for team collaboration, but in our use case we don't use Calm for that. We are not that big. We create the machines or the application; it's not that we deploy services so that another service can deploy their machines. We are still centralized, in that sense. With Calm, you can do this: With the templates, the services that need new VMs can make their own VMs, but we do not have this requirement for now. It's only used by the IT team here, which consists of 30 people.
What needs improvement?
As I mentioned, we use now CaaS from SUSE; it's SUSE's Kubernetes. But it's now changing. They have bought Rancher and I think that CaaS will be replaced by Rancher. So currently, to manage a Kubernetes cluster we have SUSE. But with Karbon we can manage Kubernetes with Calm. But I don't don't know how much we can do with Calm there. There could be room for improvement, although I'm not entirely sure. It's on our agenda to look into Karbon in relation to Calm and what we can do with them together. I don't know how deeply they are integrated. It's not necessarily something that is wrong.
Karbon is a new product. It's been around for about two years. The integration is growing. Last year is when it started working with Calm. It's more a concept still. My wish is that it will really be supported, but I cannot say for sure.
Again, I'm not saying something is wrong here. I think it's a very good platform, but there is always room of improvement.
For how long have I used the solution?
I have been using Nutanix Calm since last year. We started in 2018 with a proof of concept to go to a hyper-converged platform, and then we chose Nutanix.
What do I think about the stability of the solution?
The stability of Calm is very good. We have not had problems. We are enhancing our clusters now a lot because we did a proof of concept for two years and last year we went into production. We are really happy with the platform and we are really accelerating and enhancing it.
What do I think about the scalability of the solution?
We are a company with 700 employees. In Nutanix's world, we are not a big player. I don't think that we are ever going to push the boundaries.
We are also using Nutanix Files cluster. We are also planning to go with Era, which is a SQL management platform on Nutanix. It's really that Nutanix is providing a platform strategy for us. We are replacing all the other virtualization infrastructure that we have with Nutanix.
How are customer service and technical support?
Nutanix technical support is great. It's very fast. In the beginning we had an issue and they were very quick. The support team from Nutanix, compared to others, is amazing. They provide help really quickly. Support is really one of Nutanix's strengths.
Which solution did I use previously and why did I switch?
We had some templates in XenServer, but they were more a type of predefined image so that when you installed it helped start the machine. We also had Salt scripting, but we didn't have tools to manage them. We are not a big company. We had something like 500 virtual machines and we had templating tools and a lot of manual tasks. So things were semi-automated. We had images for certain applications, but when setting up the machine, we had to manually finish the setup.
One of the drivers for us to go to a hyper-converged system was that we had a 3PAR SAN which went out-of-support. So we had to make a decision about whether to buy a new SAN or to go with hyper-converged where you can grow with the need. And this became one of our preconditions. We wanted a system that does not use traditional SAN. We liked the idea of hyper-converged.
We bought a little machine and did a PoC to see how Nutanix works. We already knew it was a good platform because we had heard good things about it. When we tested it, it was very good and very fast and fulfilled all our needs. That made the decision for us, that it was the right platform. It became a part of our company strategy.
It was a good decision for us because now we can also replicate the whole cluster to the big cloud providers. You can have a Nutanix environment on all the three of the big ones. That means that we can buy a Nutanix cluster on Azure or Amazon cloud, for example. Then we replicate our cluster to that cluster in the cloud, and then we can switch over. With Nutanix, we can easily deploy a virtual machine in the cloud, but then we are using the cloud provider's functionality. But now Amazon, Google, and Azure make it possible to rent a Nutanix cluster. So if we replicate, and an airplane crashes into our building, we can switch over to the cloud. For us, that was also a statement that we were really going with a good platform. In Switzerland, a lot of big companies are using Nutanix now, well-known companies that are going hyper-converged.
How was the initial setup?
For me, the initial setup of Calm was straightforward. It comes with Prism Central and Prism Central is a one-click installation, and then you have Calm. It's really easy. The whole Nutanix platform is really easy to manage and to update. When you have Prism Central, you have Calm already. You must buy the license for the blueprints, but it comes with Prism Central.
If you need cluster management, if you have more than one Nutanix cluster, you need Nutanix Prism Central and with Prism Central you have Calm.
Our deployment strategy is "one-at-a-time." We touch one system and make blueprints and then we go on to the next system. We migrate machines to Nutanix without a blueprint, but the goal is that—even though we have a lot of virtual machines and use cases, and this is an ongoing process—all the new projects, as well as when we touch an old project, will go over to a Calm blueprint, to make life easier. You cannot make that shift in one day.
Our overall strategy is to have Calm as a central tool to deploy virtual machines, with a requirement that nobody manually create virtual machines. There should be a blueprint first.
There are times when it might not make sense, if you need just one machine for a particular use. It could be more work to make the blueprint. But I think it's worth making even these little machines as a blueprint, so that you can always create this machine everywhere, including the cloud, without documentation. And that's another point. As you know, when you write documentation, as soon as you're finished it's already old because things are changing.
What was our ROI?
We are still building our infrastructure, so it's early for us to look at return on investment. But there will be a return on our investment because we are not buying another SAN. We have saved a lot of money, because the SAN system is very expensive and also requires very expensive switches. So we are definitely ahead there.
Also, we had a lot of XenServers on hosts, and going with Nutanix allowed us to reduce the number of hosts. The new system is very performant and we don't need as much hardware to get the same performance.
In addition, although it has nothing to do with Calm, Nutanix helps by giving us a good overview of what is oversized or undersized. We can look at it and see, "Oh, this machine may be underused or overused," and we can free up resources. This is also an ongoing process. We see that a lot of machines are oversized and we can make them smaller. We save resources for other machines that way. But that part is Nutanix itself, through Prism Central.
What's my experience with pricing, setup cost, and licensing?
Calm comes with Prism Central but you enable features by buying the license for them. You buy by the blueprint, how many blueprints you need to manage.
Which other solutions did I evaluate?
We also looked at HPE. We compared Nutanix with that solution. We decided then to go for Nutanix and do a proof of concept. The HPE solution was more limited in the nodes it could handle.
We work really closely with HPE. All our servers are from HPE. So HPE proposed a solution to us, But when we compared it by doing a SWOT analysis, part of our consideration was that Nutanix is a newer platform. It empowers a lot of things. It's a different technology.
What other advice do I have?
My advice is "use it." To use Calm, the precondition is that you have Nutanix. To me it doesn't make sense to have Nutanix on-premise and then not use Calm. Then you would have to use SaltStack or Chef or whatever other management software exists for managing virtual machines or physical machines. If you go with Nutanix, it makes really sense to use Calm.
SaltStack and Ansible are also good, but it doesn't make sense to use them when you have Calm. With Nutanix you have one platform where you can manage everything. Calm gives you a lot of possibilities because you can script and easily integrate and control the whole Nutanix cluster with APIs. And you can easily integrate other services because you have the ability to call Python scripts very easily.
For us, it was very easy because we didn't have a lot of existing scripts. Other companies that have a lot of Salt scripts or a lot of Ansible scripts have to recreate them in some way. So we were in a good situation.
We now have 14 blueprint templates, and still growing. We are coming from the Citrix XenServer platform. We are not automatically creating a blueprint. It's ongoing. We had a lot of virtual machines on the Xen platform, and we have moved them over, but we don't automatically have a blueprint when we do. You must create the blueprints. We do them one-by-one. When we touch a system again, we create the blueprint for it. That way we can scale out, scale in, and make test systems.
There is a template for creating a machine, and then you manage that machine with this template. But when you have machines from another platform, like the XenServer virtualization platform, you can move it over, because Nutanix is also a virtualization platform for running VMs. But then you don't automatically have a blueprint, so you have to start a new project to make these blueprints. The strategy is that we will have all the code for our infrastructure so that we can build all our system out of blueprints.
I cannot say Calm is providing centralized control of all our applications because we have some legacy systems. We have IBM iSeries, which is another technology. But with Calm we can centralize all our x86 machines.
It's still early time and there is room for improvement. I give Calm a nine out of 10. I cannot give it a 10 because other platforms are also really good. Ansible and SaltStack are also powerful. It's more an issue of strategy and the fact that it is very easy to use. It's not a complex tool. They make it easy to use. Other frameworks are more complex to use, but may also be more powerful. But for our purposes, it fits exactly what we need. We haven't been blocked from doing anything we need to do with Calm. We haven't had any showstoppers.
Compared with other tools, Calm is newer and the scope of what you can do with it is still growing. They improve things. They make it easier to handle.
Which deployment model are you using for this solution?