MacStadium Coffee Chat - VMware on Mac

See how your team can transition off of VMware seamlessly

Need to migrate off VMware but worried about the transition?

You’re not alone. Many companies still need to switch to an alternative solution. We know change is hard, but eventually, VMware will stop working for your team. You need a transition plan to continue business as usual. Tune in to hear about:

• Workflow changes you can expect, and how to manage each
• Typical transition experiences and timelines
• Potential VMware alternatives
• How to have the conversation with your management team

MacStadium has already helped many customers make the transition from VMware. During this Coffee Chat we will give you the tools to create a smooth transition plan for your team.

Listen to the Coffee Chat (~18 minutes)

Rather read the chat? Check it out below:

Introduction

Hi everyone thank you for joining us for this session of MacStadium's Coffee Chat. Today we have MacStadium CTO Chris Chapman and CPO Jason Davis here to talk about the future of VMware on Mac. I will now turn it over to them to introduce themselves and kick things off.

Chris Chapman:

Hi everybody. My name is Chris as she mentioned I'm the CTO. I've worked a lot with our software and virtualization platforms at MacStadium, so have a little bit of subject matter expertise in this area, and also with me is Jason who is our CPO.

Jason Davis:

Hello, I'm Jason Davis, Chief Product Officer here at MacStadium. I've had the pleasure of working with a number of our customers who have been working with virtualization technologies, both on VMware and other technologies and really excited to talk to you today about some of the experiences we've seen.

3 Things you NEED to know

Jason:

So let's start off with the most recent experience which is what's going on in the market with VMware and Mac. There are three key things that you really should know. One, VMware on Mac has been in support for Intel only and that support for new operating systems is ending. So, for those that are still running on VMware on Intel, a new operating system support certification is going to stop. The second thing that you should be aware of is that Broadcom has recently purchased VMware and is updating and changing the way that licensing is done. This is really resulting in a lot of increased costs across the board and something people really do need to be aware of. Number three, this has been a fact for a while. VMware has not moved on to the new Apple silicon technology, which has been out for over four years now. This is the modern way that people do development on Mac, and for this reason, folks have been looking for alternatives to VMware on Mac for a number of years now. I'm going to turn it over to Chris for a moment to talk about some of the options that people have now if you're still running on VMware and we'll talk a little bit about some of the considerations that you want to take as you're moving between different platforms.

Choose the Option that is right for YOU

Chris:

Thanks, Jason. Kind of breaking it down, you've got sort of two choices: You can push this decision off as long as possible which means you're in a sustain mode, you're going to defer. With that comes obviously the burden of maintaining your VMware, as Jason mentioned support is going to be very limited and getting more limited. Licensing is going to definitely go up so you're going to pay more for what you're doing today. And then ultimately as the life cycle ends, you're going to be providing your bug and security management. You're going to lose the ability for anybody to really help you maintain an SLA because again there's no support. And ultimately it's going to have to drive you back into the older Mac OSs because there's just no development for what's coming now and next. We hope that alone is enough to sort of give you the idea that what we think the right choice is here is that you consider migrating and adopting a new stack, for the Apple ecosystem it's the right thing to do. And because of that, there's sort of different ways you can attack this problem. You can do it yourself, you can use bare metal, and you can sort of try to find an analog to the virtualization platform. With that, let's talk about DIY for a minute.

You know there's a bunch of technologies available, you can download them as software. There are things like parallels which is more of a desktop experience. You have open-source projects like UTM and Tart. Tart actually bridges open-source and paid, paid solutions like Anka. Then you have a combination of you know Apple puts out the hypervisor framework spec and everything else you can sort of try to roll your own and build your own, but with that comes the need for compute, the need for network, and the need for storage. Which means that burden is on you to provide your own Mac hardware, figure out your own network situation around that, and figure out how to bridge that with the storage that you need. And again, depending on your scale and complexity, that can be straightforward or that can be immensely complex.

Jason:

Actually, Chris, that's interesting. We're working with number of customers, some of them are starting to move towards DIY. Some that are trying it and coming back to the cloud. I find one of the most interesting things is how Mac is not built for a data center. So running a fair amount to compute requires kind of bespoke mechanisms to get it set up if you want to run it in a data center. That's one of the things I found really attractive at MacStadium is that expertise that we've built up hosting Macs in the cloud, making sure that the racks are as dense as possible, cool is possible Etc. So definitely a lot of complexity and not for the faint of heart for sure.

Chris:

Yeah, and that's a great point. It is probably a bigger bite than you may initially think because it is relatively easy to download these tools on a single Mac and get started. And they're designed to be easy and help you get your head around virtualization which is a great start, but the scale and buildup of those quickly becomes a challenge. And it's not a knock against any of the tools, it's just what's involved in solving the technical problem to do that. As Jason mentioned, MacStadium's kind of, it sort of goes unnoticed when you use the service but there's a heck of a lot of engineering behind what's going on under the covers. Which leads us into Bare Metal Solutions.

AWS is obviously the player everybody knows in the space on the non-Mac side of things. They do have an ec2 Mac offering. MacStadium, who we are, obviously we focus wholly and totally on Mac all day every day. Then there are third-party providers with varying degrees of complexity that provide sort of hosting or some of them service some of their offerings as well. But the advantage here is at least from the compute network and storage side of things, the bare metal provider takes that burden from you so you have less of the problem to solve. You're still dealing with a physical host or a fleet of physical hosts though at that point, so your CI pipelines, your machine management, your machine state management, and those types of things, those are still going to be on you. You're going to have to bring your own automation about how you handle and engage this fleet. And as a former VMware person used to using sort of ephemeral or flexible scalable virtual overlays, physical is a bit of a jarring experience because it takes away a lot of that capability. Bare metal is just a bit of a different animal but definitely a less complex animal than say DIY.

So the last piece is trying to stick with virtualization or an advanced form of virtualization and convert completely. That's where we think Orka really has some power and capability you should look into. Again, it gives you the capability and performance, and sort of the tool ecosystem that you're used to in a virtual environment, but it adds a little bit more around it. And we'll we'll talk more about that in a minute. In that case, it's a virtualization to virtualization conversion versus sort of a a system rebuild.

Technical Considerations

Getting into that, regardless of which direction you pick, want to help you plan and gear toward whatever you do. Obviously if you sustain, that's do nothing but eventually that runs out. And when that day comes, whether it's DIY bare metal or virtualization, here's some of the things you need to sort of dive down on.

You're using VMDK images on VMware, very standard, well-known, well understood. Those will be converted to either another virtual form factor if it's something open source, especially on Apple ARM. You're going to use their hypervisor framework and that uses a totally different image type. VM tools and VMware also come with VM tool integration. Those tools are what live inside the VM. They provide tighter integration to physical interfaces, to storage, to networking. They overall amp up the performance and make it act like more like a real physical machine. Those tools do not convert and do not play well in any other format so they'll need to be removed and you'll need to make sure if you're doing anything that specifically relies on the benefits of those tools that you're aware of it.

The next thing you need to do is sort of look at your workflow. Are you doing anything that's VMware management system or VMware tool dependent in your workflow? For example, if you're leveraging the capabilities of say VMotion and Storage Motion to move your fleet around or manage your images, that's going to have to change. If you're doing something around load balancing where you're letting sort of the built-in VMware post optimization where it balances the resources for you under the covers, you may move to a new system and start realizing you've been over subscribing hosts or you haven't been managing the resources as carefully or you need to learn a new way of managing that resource load balance. And then thick and thin provisioning, a lot of people use thin provisioning again that gives you the ability to over subscribe your hosts. That has pros and cons depending on the workload. Then thick provisioning where they just treat VMs as sort of long lived things that are always available to target. Those types of workflows affect what you do in the next life.

Automation changes are another big piece of this puzzle. For sure what's going to change is your interface to things and around that Powershell scripts that are driving your VMware platform, those are going to go away or need to be converted to the scripting that fits the new platform or target that you choose. And then from a CI perspective, whatever plugin, even if it's something very popular like Jenkins, or if you've written your own API integration, you're going to need to change those interfaces out. They may be very simple, it may just be a matter of swapping out the plugin and filling in the right parameters. The API integration layer could get more complex depending on how deeply you've tied together.

All of this adds up to the planning and transition involved in this. It's not a magic switch, there isn't just a quick conversion tool because there are multiple dimensions to this problem. You need to evaluate the real timing for your solution and your scale, and that could be weeks, that could be months. It really depends on the size of the team and the complexity and the integration level. So it's not to make this daunting or terrifying, it's to get realistic about what is it going to take to get from this platform to the next platform, and how much work is involved. Because that is important in planning and setting out how you're going to do this. Also one thing to consider there that adds into that is testing. You're going from a well-known platform that you understand to something new. Even if you start to understand something new, you're going to want to test side by side at the atomic level with VMs, so that you understand your just general build behavior, performance behavior, and side-by-side behavior.

And then leveling up from that, how do your workflows work? How much end to end time is involved in each system? How do you update and manage those systems when you need to make changes or when you need to grow or reduce? All of these things add up, right? And again, no matter what direction you're going in, I would strongly urge you to consider all these factors to figure out what your real plan is going to be. With that, I kind of want to dive into a more prescriptive thing that we understand at MacStadium which is the Orka platform itself. I'll let Jason take it from here to talk to you about what some of our customers have done.

Thumbtack Case Study

Jason:

Thanks, Chris. I'm going to talk a bit about Thumbtack, one of our earlier customers with Orka, they have been with us for quite a while as they were transitioning away from VMware on Intel onto the new Apple Silicon. Thumbtack, if you're not familiar with them, they provide a Marketplace for homeowners and repair and renovation professionals like plumbers and painters to and electricians to connect and be able to engage on with each other on projects. As you might imagine, iOS apps are really important to their business. In fact, they have two different apps that they have to support. And one of the reasons that they looked to make the move from VMware Intel to MacStadium's Orka and Apple Silicon, was the demand that was being put on their build systems. Every pull request, it's run through there. Every integration test. In addition to every CI build, as well as all of their automation tests. So they were seeing the need to get more performance and the ability to quickly configure environments as they need them depending on what type of workload was happening on the servers.

Thumbtack made the move from VMware to Orka and saw an increase of essentially double the amount of throughput that they could put on the system. Their build times were cut in half, and they were able to support more workloads on the same level of the same number of hardware nodes. Great case study and indicative example of the type of improvements people see moving to the newest technology, which again, VMware is not supporting. I think now it makes sense for Chris to talk a bit about Orka and the new version of the product, the 3.0 version that we released about six months ago and some of the highlights and capabilities there.

Orka Highlights

Chris:

Thanks, Jason. Yeah let's get into a little bit about what Orka is and some of the benefits of it. In general, Orka is a CI-centric orchestration platform for Mac. So again, sort of different from the broad use purpose of VMware. Orka is specifically targeted A - to run Mac and B - to interface with customers who need to do build and development. Some of the highlights in the latest version are that we've gone k8's native. So what that does is not only does it allow Orka to operate on Kubernetes, which it has for a while, but it actually unlocks the tooling and interfacing of the Kubernetes ecosystem because Orka now lives in Kubernetes in a more native way and that lets you integrate with all that tooling, lets you use kubectl, let's you use tools like that.

We have added a pretty significant compression and wrapping technology, not just with the OCI wrapper, but with some compression on the Apple side as well. Basically it means you can thin provision these containers and distribute them across your fleet, and with OCI and you can now manage them with registries that most dev teams are very very familiar with, which helps with versioning, which helps with fleet management, and on and on.

One other big capability is that we've integrated SSO. Identity management and single sign-on are critically important. You want to be able to have ARBAC capability that's built into Orka exposed out to your larger user base and development team for being able to segment by team and role and capability, and that's all tightly integrated into the product now.

See what customers say about Orka

Jason:

Yeah we've seen really great uptake from our existing customer base. I think 90% are now on the newest version and being able to take advantage of those improvements, especially around performance, spin up time, the image size, and so on. Many customers have lots of nice things to say about Orka. We've really appreciated working with them and they appreciate the technology. It's helping them get their job done. We talked a bit about Thumbtack already and how they're now prepared for future version of Apple silicon.

AppDynamics is another great example and might be one similar to large organizations out there, where you already have ephemeral build systems for Windows and Linux, and it's time for Mac to join the party, and Orka allows that to happen. Glovo is another example of an iOS app. Glovo does food delivery services. You can only imagine, they have multiple check-ins happening, multiple builds happening, and integrating with Jenkins is very simple for them and easy for them to get up and running and transferred over onto the Orka platform for running their builds.

Not only these customers, but we're working with dozens of customers, probably in the same industry as some of you. Tools that you're familiar with like Homebrew to brands like Disney and Wayfair and Cisco. They're all doing work with us - Apple virtualization with Orka and bare metal running through MacStadium. We look forward to trying to work with you as well.

Chris:

Thanks everybody, I hope you got some good information out of this and if you'd like to learn more and work with us, we'd love to work with you. You can scan the QR code here for more information.