Broadcom promised to reform VMware so it enables better hybrid clouds. Will it deliver?

Adopting – or increasing the use of – a proprietary computing architecture like IBM's POWER is a very niche thing to do in 2024. Yet in June, Big Blue suggested doing just that: employing its single-supplier stack as a replacement for the VMware stack. IBM cited changes Broadcom brought to the virtualization giant as a reason …

[Charlie Clark]

Anyone who stays is an idiot

VMWare might think they've got away with this but as soon as one customer moves, others will follow. It takes a while to come up with a migration plan, test, validate and then do it. But if it works, everyone else will want to do it and it's a part of the stack where there's a lot of exchange between developers and system admins.

[Gene Cash]

"Where is the value in moving to an equivalent?"

Isn't that a rather stupid question when the equivalent is half the price? Am I missing something here?

[Charlie Clark]

Re: "Where is the value in moving to an equivalent?"

You have to consider the costs of migration. VMWare is obviously hoping this is going to keep more customers with them longer.

[Anonymous Coward]

Re: "Where is the value in moving to an equivalent?"

Be warned, PROXMOX isn't a silver bullet solution. It's fragile with some frustrating bugs that haunt me on my home lab. I run Enterprise level hosts and storage at home, by the way. When it works, it works well and has some decent functionality, but being forced to reboot a host because a vm refuses to turn on isn't fun. No cold migrations between hosts, only works with vm's powered on. Fun when you need to move around DR servers that cannot be on without disrupting the production env.

XCP-NG is next on my list, but colleagues who tried it have trashed that as well for being inflexible and almost as expensive as VMWare.

Been hands on for three months now and I wouldn't use PROXMOX in a production environment. Test / DEV maybe, on a bet. Production? No way in hell.

[Anonymous Coward]

Re: "Where is the value in moving to an equivalent?"

Agreed on the general sentiment (prod vs. test), but I think it's worth remembering there are variations and levels of both.

That is, "prod" at some places might barely qualify as "test" in others, plus the reverse situation. Horses for courses.

Various analysis articles of the Broacom+VMware strategy, both pre- and post- buyout, theorized that Broadcom's plan is likely to keep the big-dollar big-contract "enterprise" customers locked-in, and maybe best-effort the rest. Reasoning is simple enough: the big customers are higher revenue, and probably less able to migrate away from VMware -- hard to change course on the Titanic, and all that.

Whereas the smaller outfits might be better equipped to migrate, more tolerant of test/dev foibles, and thus more willing to try other solutions. E.g. I ran a dev lab on a handful of vanilla KVM-qemu hypervisors, on boring last-year's model bare metal, no bells or whistles. The engineers carried on with it fine -- they were only interacting with VM's we spun up for them anyway -- and saved us a little money to offset some of the very expensive vmware license packs. Would I have wanted to run a prod database with failover shared-storage etc. that way? Not a bit. But for mostly ephemeral and generic VM's for developer types to (ab)use until they needed a new one or a different OS etc., it worked out.

The prod stuff stayed on VMware, and it was fine too. Though I wouldn't expect it'll stay there forever.

And that's the thing: I don't expect this is a long-term winning strategy for Broadcom, and I suspect they know that. So Broadcom are wringing out whatever profits they can get for as long as they can, and maybe VMware eventually follows some of the other Broadcom acquisitions.

[shfy70]

Re: "Where is the value in moving to an equivalent?"

Agree with your sentiment, maybe the only benefit for big org such as BOA/CHASE is the $$ per license is 20% cheaper versus mid size org (having 100 ESXi), but the sheer size of BOA/CHASE with >10,000 ESXi make the lock-i even stronger, BOA has a motivation to migrate away from VMware too, otherwise their CIO may lose his/her job, the most likely scenario is almost all big customer will sign up for VMware license for 3y and plan to migrate to AWS/GCP or some smaller sites to Nutanix ( they are shady as hell and will match up price increase too, remember they are American company on NASDAQ as well)

Bottom line: never trust any large IT shop, NEVER! Get rid off your VMware expertise ASAP

[sedregj]

Re: "Where is the value in moving to an equivalent?"

"No cold migrations between hosts, only works with vm's powered on"

I've just created a VM on one host and moved it to another without it being switched on. The GUI describes the task as an "offline" migration.

[Jadith]

VMware by Broadcom rejects accusations it has increased prices, suggesting rather that the market misunderstands its new bundles are cheaper than the previous combined cost of the products they contain, and that support – previously a separate purchase – is included in its subscriptions.

This is such an infuriating insult to here from Boradcom. It is essentially accusing people of being too stupid to know what they want/need from VMware. PAYING FOR THINGS WE DON"T WANT OR NEED MEANS OUR COSTS ARE GOING UP. Just because you bundle the product everyone wants with some crap that noone wants does not mean you are adding value. It means you are adding shit. You are enshitifying your product. The fact that Broadcom just keeps repeating this over and over doesn't make it anywhere near true, unless you are some delusional stockholder, perhaps. Or maybe someone desperately clinging to the delusion that you haven't taken a garunteed money maker and enshitified so noone wants it. (Seriously, if they would have bought VMware and just let it alone, it likely would have just output value to their portfolio for years, even decades.)

Also, adding a support package hasn't added value to anything in many years now. Unless you are platinum toilet paper level, you would be lucky to get a 'have you tried turning it off an on again'. (reboot procedures take too much call time nowadays.)

HPE has created a virtualization product that looks a lot like a direct rival to VMware.

At this point, anyone that uses anything from HP deserves what they get. (Yes, technically HPE is a separate company, but they are the same in practice)