Ditching VMware over the Broadcom buy? Here are some of your options Broadcom has yet to close the deal on taking over VMware, but the industry is already awash with speculation and analysis as to how the event could impact the cloud giant's product availability and pricing. If Broadcom's track record and stated strategy tell us anything, we could soon see VMware refocus its efforts on its top …
Indeed - very much not pointless. It means that firmware updates can be done with no system downtime (although it's sensible to do this outside of normal working hours). It also allows replacement of physical hosts without downtime under some circumstances (we've done this recently with no issues).
The Xen hypervisor is still alive under the auspices of XCP-ng, and combined with Xen Orchestra appears to be pretty much at feature parity with our current small VMware + vCenter server deployment. We'll be evaluating it as a replacement over the next few months as we're due a hardware refresh.
I ditched VirtualBox due to Oracle's inability to keep audio devices like the speakers and microphone working, which is rather necessary for virtual meetings. Every new release would break audio and it'd be 3 months before it was fixed.
I was a little worried about how difficult it would be, but I installed KVM and it turned out to be a doddle. It Just Worked once I figured out the virtual drivers for things like ethernet and displays.
Red Hat Enterprise Virtualisation is based on the RH-incubated oVirt project - an open-source management layer around multi-host KVM. In my experience though it had many rough edges and is very opinionated on how you setup your cluster, not reacting well to mixtures of hardware types and storage back-ends.
Also there's the re-incarnation of XenServer... XCP-ng with Xen Orchestra. Never used it, looks promising. I was a fan of proper original Xen many years ago.
I have not looked at what's happening in the OpenStack world for some time. Last I looked it was mostly bare-bones platform-as-a-service type projects rather than a fully integrated virtualisation platform.
Proxmox is worth looking at for smaller clusters and is more polished and flexible than oVirt/RHEV. If I was looking for a budget homelab or other cheap setup I would go with it.
Thanks for the tip on OpenNebula ^ I'd never heard of it and will have to check it out.
Honestly though when you compare the flexibility of cluster hardware support, mixture of storage types, live storage and compute migration... and vSAN... you have to be mad to look at anything other than VMware vSphere for small biz and larger org non-cloud/on-prem. I don't even particularly like vSphere but for the most part it "just works™" moreso than any competitors. I have been paying for VMUG EvalExpress license for some years for my homelab tinkering.
We (education) run motley mix of workloads as our industry standardized on EVERY damn platform and the individual vendors generally don't GAF. At least dealing with Virtzilla meant most of the vendors had at least heard of it. might have tested on it, and one or two would just ship you a virtual appliance.
Hyper-v is the next logical choice, and is dirt cheap and basically included with our education licenses for server. It is also like moving backwards in time as it suffers under the tyranny of all of the M$ server tools. VMwares decade long GUI migration failure still wasn't enough for them to play catch up. I'd have to hit the command line enough on either that it is personally annoying but moot to me. However, this isn't a one man shop and my boss and co-worker need management tools designed for actual humans.
Past that turns into support cowboy country real fast, especially when you need a mix of windows, linux, and possibly some crusty OSX stuff. It wasn't pretty, but Fusion+esxi got the job done, and the critical parts of the tools mostly talked to each other(except the damn update manager). Hyper-v by itself can't really replace that, and a totally mixed hypervisor/baremetal mess isn't great either.
But things like M1-M2 CPU support, Vmware Fusion, etc. are exactly the sort of thing Broadcom will likely axe once it starts swinging, and waiting till they've axed the platform isn't a good career move for me. So I have been testing a migration to a mixed Hyper-V environment, just in case.
Do have a look at Proxmox. It's robust, easy to manage from both GUI & CLI and frankly just seems to work. Converting the VM's can be a bit of a nuisance but it's a one shot deal - once it's done, it's done.
You get full clustering, failover, HA, backup and even distributed storage (if you want it, we didn't and went with a traditional SAN architecture). What's more, a single web-based GUI manages all the servers in the cluster. For us at least it's more than a match for Hyper-V and easily stands up against VMware.
And it doesn't cost a bloody fortune either!
This post has been deleted by its author
As someone working in VMware PSO, and actually it’s the best job I’ve ever had, you have to wonder if the focus on the top 600 customers strategy will come out like it always has done for Broadcom, I’ve no reason to believe it shouldn’t.
I think there’s potentially going to be a huge opportunity for helping customers get out of whatever Broadcom inflicts on them. Proxmox seems an obvious choice for the smaller customers, but the mid size ones are going to want some more support. I guess a lot of those customers are going to have their migration to the cloud accelerated in one way or the other, or perhaps Broadcom are relying on them moving at the snails pace customers often trundle along at while they argue about internal processes.
It’s a very strange atmosphere inside VMware right now. Management telling people to carry on as if nothing has happened, with the air taken out of a lot of people. It’s a shame, some of the brightest people I’ve ever worked with are here and it would have been nice to have more time.
The part of my job I really enjoy is helping deliver things to customers, working with them to solve problems and getting them into that new system, seeing them migrate workloads to the cloud for example, even fixing old shit they should have sorted years ago. I guess I’ll just keep plodding along enjoying what I’m doing until the axe of Broadcom falls on myself, I think we’re too low margin for their interests.
Sounds a lot like this Onion article.
Good luck to you and your colleagues!
"And when it comes to independent software vendor (ISV) workloads, many are already validated to run on Docker or in Kubernetes clusters."
Phfffff, I'm still waiting for that list. There might be like 5 message buses or service status engines built on open source that major vendors have ported to containers. But 90% of large enterprise workloads are still an ISO-delivered WS or RHEL VM. And don't get me started on the long tail of dead-developer weirdware out there, esp. the ones that snooty doctors and clinicians "have to have".
If the deal goes on and Broadcom leaves only 1500 customers on support, everyone else that's "enterprise" will run out their support while finding a SaaS replacement or moving the VM to Azure. For those involved this will be a bigger boom than it will be for Broadcom.
Openstack is still in the running and oVirt is a viable option even if RHEV is dead. ocp-virt is NOT enterprise grade and meant to enable migrating VMs before containerization. Few complex apps running on dedicated VMs before moving to micro services.
VMware is slowly dying and Broadcom might drive that final nail in the coffin.
I've got a dev background and I've been in IT management too long now...so need a sanity check. We've got about 2,000 VMs in production with VMware. A mix of HPE and Dell servers, mostly, along with NetApp, Pure and EMC storage. We've got a lot of Oracle and an even mix of RH and Windows.
One of our techs said it'd take 2 years to get off VMware.
He also said Nutanix and Microsoft won't cut it. Does that sound right to you? Seems a little far-fetched to me.
The time estimate depends on the process and the resources involved. Our team can very probably do it in about one year with high guarantees (been doing migrations for over 12 years now), but the actual mileage may vary ;)
The alternative I would really suggest looking at is XCP-ng+Xen-Orchestra (and maybe even with XOSAN if really evaluating Nutanix, too). Please feel free to contact me directly if you require more details (or fancy a free consultation on migration procedures/support)
Disclaimer: been using Xen-based virtualization almost exclusively since 2006, XenServer since late 2009, XCP-ng since 2018... and have extensive experience with several versions of VMware, some with Hyper-V and Proxmox and exposure to Nutanix. IME XenServer/XCP is the most robust and flexible platform of them all: We could always recover from "disasters" (caused by other parts of the stack failing, mostly due to sheer abandonment) and grave human error in relatively short time and with no data loss. HA implementation is quite robust, too.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
