If you want to run Windows VMs with all the security features enabled, xcp-ng / xen is a non-starter as it's been completely broken for years with no fix in sight:
Proxmox / KVM is the only alternative to ESXi that can support all of Windows security mitigations such as Core Isolation, which requires nested virtualization.
It must be possible to milk this "dumb" investor money - which directly destroys the bought goods in the long run. Basically you create two teams. Team A develops a business version of a open source-good. Team B works on that open Source good.
Team A gets acquired. The product destroyed. Team B now founds a company becoming Team A, while Team A forks the open source product and keeps the game going. Its a ton of migrations, but you could get wallstreet to donate to open source as a business strategy.
I'm not sure this is appealing to wallstreet at all but this just describes the appeal of open source as a customer to me. VMware is a just perfect case example, they're literally firing paying customers and destroying a product lots of people are using to make a quick buck. The company and the product will ultimately be destroyed regardless of their underlying value but not before they make bank.
But if Vmware was open source they never would have bought it. The leverage being closed provides is the whole point.
There is legacy technology corporation A - with a bunch of enterprise customers, eventually they:
- reach a plateau of sales /growth
- feature development slows down
- cost getting cut across the board
The core product team eventually just leaves the company to use all their internal knowledge to create a startup B that competes with A and delivers next-gen technology in the same space. They get traction, and funding and outgrow company A eventually.
100$-200$ per VM is a fair price I think. Most of the revenue generating on-prem orgs can afford it. If you think it is expensive then you are not the target segment for VMware.
The fallacy here is that you’re ignoring the pipeline. The next generation of infrastructure admins are either running on all cloud, or running on something they can easily access like any KVM based system (like proxmox). VMware is instantly a legacy system because anyone who doesn’t already run it isn’t going to bother learning it.
Revenue generating or not, the free platforms are will generate more revenue than VMware at such and exorbitant cost.
"Can afford" and "expensive" are distractions - orgs don't buy things just because they have the money to so they find the most expensive option to burn it on. The product has to justify the cost vs other options, not just be feasible to purchase.
5 Years ago VMWare support was, in my opinion, absolutely gutter.
Calling in for any questions was met with a support tech who was actively hostile and tried their best to find some way to disqualify our ability to get support. The first trick was always Hardware Compatibility List ("HCL") followed by interrogating us about software versions. Before I get the cries of pity about all the dumb users who waste their time by not upgrading, we probably opened support cases twice weekly - and the same verification interrogation happened.
We'd routinely find situations where the stated support for SANs wasn't what it said, etc.
I'd have said it could not get worse, but reading what Broadcom is doing, I am convinced otherwise.
Anyone using VMWare for business critical things is likely in for both a huge pocket book hit and an even larger support headache.
I'll just add that upgrading cam itself be the problem- having an update come out that isn't yet hcled with your hba probably is fine, until you call support.
I vaguely remember being a very early user of VMWare, although I have long lost my license for it from the 1990s. I could have sworn there was a clause in there that early adopters had lifetime licenses and support.
Did anybody else become an early adopter but kept all that license stuff, or am I not remembering it correctly?
Broadcom must have the most idiotic management from all the large tech companies. Everybody understands you buy a company to recoup the investment, streamline operations and push forward with new markets.
But do you have to increase license prices by three times within the first quarter of an acquisition? Drop all free licenses that allowed for learning of your technologies? Even professionals with years of experience of providing VMWare official training curriculum are being forced to pay for licenses that were free.
All training responsibilities, dropped with no warning on the historical training providers...
What McKinsey Consultant, advised for aliening a customer base of 20 years, within the first quarter of an acquisition?
Google, Azure and AWS Sales reps must be salivating at this opportunity.
The view of Broadcom is that VMWare has low growth potential and the route to maximum short term financial returns is to extract as much money as possible from the largest customers.
So the strategy is to squeeze the 600 largest customers until they bleed, knowing they are so entrenched they cannot easily migrate. If the next couple of thousand stuck around as well that would be nice, but not essential. The long tail of hundreds of thousands of customers paying a few dollars a month they would gladly be rid of.
The last project I was working on at a major telcom before jumping ship was a project to migrate off of VMware onto KVM systems managed by Cloud Stack, they already have a major Open Stack environment but that's only for SDN stuff now.
We're talking over 10,000 ESXi hosts in the current footprint and at least 50% of them will be off VMware by the end of 2026.
If they were smart, the people they would be focusing on squeezing are the mid-sized shops that don't have the in-house skills or top level cost focus to force a migration. The small shops can't afford to stay on vmware, the very large shops will have the in-house skills or contract to the right people to get moved, it's the mid-sized people that are most stuck.
How do you know this is their view? I don't have any strong opinion or evidence to argue, but I'm curious if this is conjecture or inference or actual knowledge?
This is literally what Broadcom does - it's just the first time most of the HN crowd has been subjected to it. See: bluecoat acquisition, Symantec acquisition.
Drop any pretense that you want to service any customer outside of the fortune 500. Then increase prices by several hundred percent because you're so imbedded in those fortune 500s, even if they want to run from you it'll take them 5 years and you can make all your money back in that window.
Perhaps you already know this, others probably do not, but Broadcom acquired CA (computer associates) in 2016. So yes, literally the Computer Associates Strategy. "This IBM mainframe racket seems amazing, how can we replicate it???"
Note: you're all going to be complaining about Redhat in the next 5 years, because IBM knows how to get blood from a stone.
If you look at its history, Broadcom is a financial acquisition company that operates in technology. I think Tan's position is: these are (mostly) mature technologies and should be horse traded like financial assets.
Got lucky with the LSI Logic design team and Google TPU's. Presently $8 bln/yr I understand. Although such serendipity increases when you're able to buy everything you see. With the exception of Qualcomm.
They want 1 in 5 VMWare customers to jump. If they don't, they'll be forcibly pushed off of the edge. Broadcom's entire growth model is to take an established player that is entrenched in massive corporations, introduce that product to the existing portfolio Broadcom offers that customer, and then jack the price on the lot. If you're not in the Fortune 500, Broadcom doesn't want to talk to you and if they keep your business at all it'll be through digital sales or third party resellers.
I am not a fan of them and didn't like working for them, but their financial results speak for themselves. You can attack the ethics of Tan's leadership certainly, but to call him an idiot ignores the fact that he's doing exactly what his shareholders want. Growing the company at obscene returns.
People like to do business with people they like, and the Juniors of today are the future CFO, CEO and CTO's. Also IT people have a very, very long memory, and aiming to become the next Oracle is not an advisable strategy.
Being the optimizer of local minima, is bad business strategy, and smart shareholders should plug for a better optimizer.
Don't disagree. I regard Tan as a (powerful) local optimization where the counter argument to broader optimizations is that they're riskier. The latter are though more fundamental. New businesses, new processes, new products are one thing. Profit-taking is another. To get new industries you need far-ranging optimizations.
I remember when Dell shitcanned a project to do something like this about 15 years ago because they realized it would lead to information hiding. Tech support loves easy to fix problems.
Amazon uses Xen which they call "Xen 2" I think because they've made tons of improvements and optimizations over the years.
Google uses kvm. They probably contribute as well although I haven't checked.
From a quick search Azure uses a private build of hyper-v which surprises me actually but it's possible.
Edit : millions of VMs doesn't mean anything, at the end of the day it's just servers each running a bunch of VM, which people have been doing for 20+ years. There is no scaling issue there, the hard part is storage, networking, orchestration, managed services, power...
I mostly agree with you.
Its the "fabric" and tools that matter.
In so far as this thread matters,
I meant to say that VMware has a lot of great tools surrounding
ESXi (which at least used to be the core of running VMs, may have changed)
In my opinion Proxmox does not have tooling like that.
So if I had written my comment better it would be what tooling
and fabric these giants use and to what extent they share the tools
and upgrades back to open source projects.
It is somewhat smart capitalism, drive up the prices extract as much money as possible as fast as you can and when product stops selling write it all off. Executives get their pay checks from options and investors see line go up for short time...
This was inevitable given the recent layoffs. Either they’re in bad financial shape or they’re sociopathically firing employees unnecessarily. Both are excellent reasons for researching the alternatives.
