We (Red Hat) are doing great business moving customers off VMware. 20,000 VMs is not an unusual amount for our larger customers. I did a 5 min lightning talk in summer about this: https://pretalx.com/devconf-cz-2024/talk/SN93LG/
It is funny because my former company was trying to move out of vmware even before the broadcome aquisition and the local sales rep pretty much told us RHEV was a dead project but at the time openshift virt wasn't ready yet.
In the end I left before broadcom bought vmware but the status at the time was "let's wait a bit more".
I don't agree with the deprecation of RHV since it's very closely modelled on VMware vCenter in terms of the concepts and user interface. So it's ideal for ex-VMware admins. But here we are. Containers & Kubernetes are more buzzword-compliant than virtual machines.
Basically, RHEV as a VMware takeout project never gained a lot of traction. With the general industry shift towards containers and Kubernetes plus the Broadcom acquisition, kubevirt in combination with a much more mature OpenShift is a lot more interesting.
Red Hat (obviously) and Microsoft (increasingly) aren't all that focused on Windows as an operating system on the server. And the desktop, in general, is about the browser with some relatively niche exceptions.
Probably, outside of VMware, you run Xen for Windows servers.
Odd way of phrasing your question but my general thought is your <10 server, no IT staff MSP customers aren't the focus of this conversation or most of the userbase here.
Looks like that tool typically targets KVM as a source. My understanding is that Red Hat's own virtualization platform is transitioning from RHEV to OpenShift. If that is true, are you actively moving customers from vmware to OpenShift via Kubevirt?
Upstream, virt-v2v supports conversions from VMware to either oVirt (RHV) or KubeVirt (OSV), and we don't plan to drop the oVirt support any time soon.
However Red Hat now only supports conversions to OSV, since RHV was deprecated (sadly).
The biggest problem is oVirt itself has not proven to be a very sustainable open source project. If oVirt dies, we'll likely remove support in v2v. (There's a great start up opportunity here, for a dull but money-making company that productizes oVirt again.)
To move VMs from RHV to OSV you can just copy the disk image since they already should have virtio drivers, qemu guest agent, and be able to boot on any qemu/KVM-based platform. I believe there's some automation for that, but it doesn't involve virt-v2v.
Thank you for taking the time to explain. I was honestly a little shocked to hear about RH moving away from RHV, but glad to hear there is some interoperability between some of these tools.
There never was a lot of interest in RHV as a like-to-like VMware alternative. But, when I left, there was quite a lot of interest (especially post-Broadcom acquisition) in a Kubernetes/cloud-native friendly VM alternative to VMware. And, yes, there are various migration tools though I was not involved in the details.
> There never was a lot of interest in RHV as a like-to-like VMware alternative
That is because it was before Broadcom lit VMware on fire. Nowadays everyone is looking for a like for like VMware alternative.
> Kubernetes/cloud-native friendly VM alternative to VMware
Cloud-native and VMs for the people used to the VMware world are antithetical. Overwhelmingly we're talking about carefully managed pets, which is exactly the opposite of cloud-native.
I agree, hence my idea about resurrecting oVirt (see up thread). Red Hat are all-in on Kubernetes, which is the wrong model for pet VMs, but a better model for gitops stuff.
I'd argue it's still essentially backward facing. Yes. There are buyers who would essentially like a cheaper VMware clone today. But I'm not at all convinced that's a good business model for a company like Red Hat to invest in going forward.
> There are buyers who would essentially like a cheaper VMware clone today. But I'm not at all convinced that's a good business model for a company like Red Hat to invest in going forward.
Yes, they're buyers stuck in the past. If RH can bring them on and take them on a modernisation journey, it could be very lucrative (even if very time consuming).
Broadcoms policy seems to be to aim for the larger customers where a 10x price rise for a suite of products isn't an issue. Charging 10x the price to 10% of your customers isn't necessarily a bad strategy for a business.
> Steve McDowell, chief analyst at NAND research, told The Register that VMware by Broadcom is “laser focused on high-revenue, high-margin business” and has priced its wares “just below the pain threshold for customers they care about.”
There is a famous story about business in the UK when some sales manager at Coca Cola raised the price by a small amount for JD Wetherspoons, one of the largest chains of cheap pubs across the country. The manager must have felt very smug about the extra revenue they'd bring in, from a small change that the client could easily absorb and could hardly moan about. What else could they do, move to Pepsi?
They did exactly that, and the sales manager was fired over losing one of the biggest clients and the loss of revenue that came with it.
One needs to be very, very careful about both the pain threshold and the "** you, I'm moving my business elsewhere" threshold.
But if Broadcom increases price 10-fold and loses half the customers by revenue over the next couple of years, that reflects in their bonus and they can then move on.
The customers might have a plan to move off vmware by 2027, but continue to pay the money until then. Come the 2027 renewal the execs that made the massive bonuses for increasing revenue have left and then when the company says they aren't renewing it's the next generation dealing with the problems, and then they just reach for the three-envelopes.
This really is on the companies that choose a single supplier. If you had two providers, then shifting doesn't take a 3 year program, it's just a policy change to prefer Provider B rather than Provider A.
It's worth bearing in mind that it's not just a financial gamble but a reputational one as well.
Chances are most of that 90% of your customer base who were previously fairly happy are now very unhappy with a forced move to another platform, so in 10 years time when they're in some other business and making purchasing decisions, someone will mention VMWare and the only thing they'll remember is that shit show where they spent 6 months migrating off it because of some ridiculous pricing change.
That’s the big problem I see for them: all of the VMware guys I know _hate_ this move and are talking about being betrayed. Even if they aren’t immediately migrating, I don’t think many people are going to be comfortable increasing their dependency on Broadcom, whereas before that was as close to a given as it gets in business.
When I worked as a SE, we'd often categorized sales reps as 'hunters' vs 'farmers' when it comes to managing the accounts. The hunters would hit their numbers, collect their bonus, move on, and leave the account in shambles for us to try and assist the new rep in recovering. This is the first time I've noticed that attitude in the company leadership. My heart goes out to our (now) Broadcom rep, as I can tell they are putting the squeeze to all contracts, to their mid-term detriment.
For the vast majority of companies, supply chain diversification represents a cost with little immediate benefit. In the long term, it can be a rational hedge against hostile supplier lock-in, but the vast majority of companies aren’t good at making long-term bets.
Apple[1], for example, has performance based RSU’s for its executives. RSU’s vest over 3-4 years and the number granted is dependent on Apple’s performance relative to the S&P.
Stock prices theoretically have future cash flows built in but markets are inefficient. In this case, if Broadcom starts losing customers in 2027 and revenue declines 5%, it would get hammered in the market at that point in time, punished much more than the 5% drop.
[1] From Apple’s SEC filing:
The value received by our named executive officers in 2022 from long-term equity incentives reflects exceptional stock price performance over the applicable vesting periods. As a result, for the applicable performance periods, performance-based RSUs vested at 200% of target for our named executive officers based on Apple’s TSR relative to other companies in the S&P 500 (“Relative TSR”), which was at:
• the 96th percentile for performance-based RSUs held by Ms. Adams, Mr. Maestri, and Mr. Williams, and
• the 99th percentile for performance-based RSUs held by Ms. O’Brien.
Mr. Cook was granted an equity award with 75% performance-based vesting and 25% time-based vesting
Theoretically, stock compensation should do that. In practice, many/most investors have a pretty short time horizon in mind as well. I'm not sure how you provide an incentive for long-term company success absent very deferred compensation which I imagine most people reading this wouldn't favor either if it applied to them. Bird in hand worth two in the bush and all that.
And most importantly some clients will not communicate that they are unhappy and just move on quietly. Especially for something like virtualization which is essentially a commodity.
I would guess the plan for the larger companies is to sell a whole suite of products around the VMware: management, ops, auditing etc, plus take very good care of their niche requirements so moving becomes a much more expensive proposition
But does it not make you more vulnerable to competitors? Losing one of your few big customers will hurt regardless, but it likely hurts less if you have a pool of SMBs to act as a buffer.
In terms of competition it's really the same. Whether you fight with a team of 1 for 5 businesses or a team of 10 for 1 business that is 50x bigger, on average as long as you have equal competition in each segment you walk away with the same. But you get very different customers, for better or worse.
A segment of a customer base is hardly a buffer, if you underserve them they will leave, and you will have wasted effort making an inclusive sales and support process to them.
That makes not losing those customers more critical: if you lose one of the 50 smaller businesses, your revenue is barely affected; if you lose that 50x whale, you’re out of a job especially given how hard it is to find another one of that size.
This shows up especially in the context of hiring: for a generation, there have been a ton of people who learn VMware at businesses of all sizes and your large shop can reliably hire them. Shutting down smaller sales affects that directly and because the smaller places are going to do something else it means that the people who move up to a bigger place are going to have direct firsthand experience contradicting what the VMware sales people are saying about how risky it is not to use their software.
My first reaction was that you're wrong, but you're probably correct in that large scale VMware installations are much larger than 3000 servers. The small side is going to be a large set of customers, probably the majority.
There are tons of VMware installation that are 3 - 20 physical server, more so than 10.000 server installations. Right now Broadcom is telling those customers, and those with 500, or 1000 servers, that they can take their business elsewhere. Broadcom is much happier to milk the top 5% of customers, compared to servicing the "bottom" 95%.
Broadcom seems to be cutting of to large a chunk of customers. They also are preventing new large customers from coming in. No one starts out with 10.000 physical servers and no one wants to switch a growing 5000 server datacenter from HyperV, KVM or Nutanix to VMware.
In this case you're probably better off thinking of these things in roughly logarithmic terms rather than absolute. By the time you're up to 20,000 servers, you are going to have roughly the same issues to solve as a 10,000 server install and a 30,000 server install. In absolute terms there may be VMWare installations that seem much bigger, but in logarithmic terms there aren't that many more classes above them, and they get rarified quickly. The same ways that you're telling this customer to get stuffed are going to be affecting even the bigger guys in a way that you can't necessarily compare the experience of this customer to the "10 VMs on one server" customers. It's risky to do that. I'm not sure there's as much differentiation in the market to be exploited as Broadcom thinks. Even their larger customers are not just going to sit there and take it... they're just going to move more slowly. But at that scale, VM migrations have a very natural granularity to them; you move them one at a time. You don't need to drop VMWare next quarter; you can just begin your migration project and move away at a pace of your choosing.
But then again, Broadcom may well be explicitly viewing this internally as "we no longer have a competitive product, there is either no way to differentiate it again or we're not willing or able to invest in what it would take, and so the game is just to milk all the money we can out of this", in which case this is pretty close to what I'd expect to see. Where I personally sit in their customer base now, VMs are a commodity and I've got no reason to pay for them anymore. That's been true for a while. I'm speaking here just as A Guy, not as an employee of anyone, but I would expect that this reality has been slowly but surely crawling up the market from where I sit at the bottom.
Hell, AT&T just settled with Broadcom. Five bucks says they stay right where they are. Hock Tan knows what he’s doing to force the big players to pay and extort them to the nines.
A less charitable (but I'd argue, likely more correct on balance of probabilities) is they're aiming for the larger customers who are so entrenched in their use of the VMWare product stack that they can't feasibly migrate without taking a huge one-time hit to their balance sheets.
i never understood, even 15 years ago, why people still use vmware when so many open source (i.e. virtualbox) systems existed at the time. The fact that Broadcom buys it out and then 10x the price on people is just even more reason to move to open source. As noted in the article, the open source solutions are 2x faster anyway, so they customers can realize a 2x increase in slices per server.
The ones staying with VMware, permanently or near term, are likely using products for which there are not good replacements. The open source alternatives can do what ESXi does sure, but its progressively more difficult to replicate VSAN (including stretched clusters) and NSX-T (with distributed firewalling) without bringing in multiple vendors that will all point at each other when there is a problem.
There are some open source options that come close to what vmware's offerings could do, but virtualbox for sure ain't that. It has an extremely limited feature set compared to other offerings in this space. For example, it is mainly intended to manage VMs on a single host. Most organizations looking at commercial offerings have multiple hosts (or racks, or even datacenters) they want to manage. Apples are much more similar to oranges than vmware is to virtualbox.
I can't speak to your buggy claim since I don't know which feature you are complaining about, and if it's part of the extensions then the "many eyes, shallow bugs" does not apply
The desktop is mostly a pretty uninteresting space in general. Virtual desktops never really took off (which VMware was into at one point) and desktop virtualization is generally a commodity for the relatively few people who use it.
not smart if you make it hard to use and learn and the next generation of technologists gain experience and familiarity with other solutions and improve them to be superior products.
Bold of you to assume Broadcom cares about long term sustainability.
From what I have seen they are just a PE firm pretending to be a tech company. They will bleed the customers who can’t move fast enough. Burn it to the ground to get every last dollar.
Unless they have a monopoly on the tech, that's not how the free market works. Raising your prices x10 without having a government mandated monopoly, is a great incentive to create competitors.
From Broadcom's perspective, that seems to make even less sense than a simple price hike that lets them shed unprofitable customers. What's the point of doing that? Some kind of sales force optimization?
Everyone now has to ask the question, should i continue paying for product that does XYZ from VMWare's competitor when i'm already getting a product that does XYZ as part of the VMWare licence.
Chances are if you're a vmware estate, those products are going to integrate better and then you only have a supplier, support portal, customer rep, etc..
Broadcom loose customers who arn't really invested in to VMWare and those who stay become even more invested in to them.
>those who stay become even more invested in to them
Maybe. Or maybe they understand the play being made as well as everyone else does, and won't be inclined to become any more dependent on an organization that will turn the screws nice and hard on anyone they know won't be able to leave them.
A sale isn't free. It requires time, headcount, and engineering+PM resources.
At Broadcom size, you want to maximize sales to F1000 customers (the kind that can and will shell 7-8 figures in TCV) as that gives you the most bang for your buck from a SalesOps perspective.
A major issue with VMWare was their sales team would chase logos at the expense of an optimal sale, so their account teams would give extreme discounts (70-80%) at barely above the margin in order to close deals.
Well, we're probably a F10 company and my project had to move off of our vmware environment because prices were too high. Even at enormous companies, these things matter, and there's still accounting for resources at the local level.
I work at an IT outsourcing company that also has some VMWare farms (though not sure if we reach 20k VMs).
What I picked up is that before the acquisition, price was mostly determined by CPUs on the hypervisor, now it's mostly by RAM.
Which means we now have to optimize hypervisor hardware to a different licensing model.
I don't think we've seen 10x price hikes, but it could possible if a provider has optimized hard for a metric that is now less relevant, and/or has difference licensing deals.
Caveat: I've picked all that up second-hand through the grapevine, I'm not working with our virtualization farms directly. So not 100% on top of what's going on.
It's the Oracle pricing model. Turn the customer upside down and keep shaking until you've got all their money. They will dress this up in different ways to different customers, but that's the model.
Having been in the field for some time and having some contacts, I must say: 100% of the companies where I get some insights are currently either already building alternatives or reviewing their options with a strong intention to leave VMware behind. And in many cases this is quite a huge task, but certainly doable with the required momentum.
I'm in small Devops team and people are pissed off. Thankfully Proxmox works just as well for us (some would say better). Even a large-scale migration seems scriptable to me.
Proxmox looks great, but unfortunately it doesn’t come with encryption oob and manually applying it comes with feature loss and other problems. ZFS vs encryption is a bad place to live. You of course need encryption for your VMs on the Enterprise.
Proxmox is a lovely piece of software, highly recommend to everyone. But I think Proxmox and vSphere target rather different scales of cluster. Proxmox starts to choke once your past 20 nodes or so. Corosync just wasn't designed for huge clusters.
I think the price hikes have made the "medium" use-case more uncertain.
I know VMWare works great for a lot of people, but it was always just a thorn in my side. Almost a decade ago my small company moved from VMWare to ganeti, and we've been really happy with it.
At the time we were running a SAN and the small VMWare Essentials license. But managing it with Windows was painful for me personally, and our HP SAN was kind of painful, every year like clockwork it would just stop processing until we had remote hands reboot it. And every year they'd tell us "Try this firmware" to shut us up for another year, until eventually the product reached EOL.
We had been running ganeti in our dev/stg cluster for years at that point (and the python.org infrastructure was using it at that time as well), so I decided to transition us to local storage running VMWare and then to ganeti. Doing local RAID array monitoring on VMWare was quite painful at the time, maybe that's gotten better?
I absolutely love that ganeti has such a good suite of command-line tools for operating. I have a small shell script that respins half our dev/stg infrastructure every other day. Our entire critical infrastructure is HA at the machine level, so we don't play around with anything like live VM moves, we just do offline moves. I used to do the drbd-based replication and live moves, but the doubling of disk writes and committed disc usage wasn't worth it for us.
Proxmox is on my radar if we needed to move away from ganeti, but so far we are looking like we'll stick with ganeti.
The latest release is almost 3 years ago, but the repo has commits a month ago. So I guess it depends on your definition of "alive". As "stable" software, I'm fine with it not having a ton of active development. It's been a real workhorse for me, I've got it running on 2 clusters of 6 physical machines each (1.4TB total RAM, 50TB disc, each).
~6 years ago I went to a local Dell dog and pony show where someone at the next table was poo-pooing on "those greybeards that run Debian and ganeti" instead of VMWare. But, this shit works...
Good to hear. VMware used to be honestly my preferred option because it handled certain hairy things well out of the box that many other solutions either ignore or require non trivial integration work (for example it's trivial to connect a VMware VM to piss-poor external network by simply attaching a VLAN - which matters when the network is a fieldbus in some OT deployment)
Ganeti I remember as comparatively harder to deploy (nothing was harder than openstack though)
So Broadcom is basically just milking VMware until it completely dies? I guess you should always consider the risk of a company buying a product you use and then doing this
VMWare has been dying a death of 1000 cuts for years. It's been passed around to now the third owner. They created an entire market and have been watching it erode away.
This is exactly the problem with optimizing metrics though. It seems like they’re getting more money from less customers but it puts them at might higher risk that those big customers will churn and tank the company which also increases as the price goes up as it will eventually hit a point where they can find something cheaper.
Yeah but how does it make any business sense for Broadcom to be doing this? They could have raised prices 100% and fewer people would flee and they'd make more money overall. 1,000% price raises are just plain extortion.
If the action of a company doesn't seem to make sense, ask yourself: could you imagine an incentive structure within the company that could lead to such actions?
Considering half the comments are “Proxmox” and “Virtualbox/Parallels” are so much better anyway, I’m guessing most commenters in here are extremely naive and not aware of the IT landscape.
This is the classic "golden goose" strategy. Why wait around all week for steady returns when you can just rip the thing open and see what you can get for the internal organs.
Mathematically if they can extract say 1.5 times the money they used to buy something in 3 years. It is a good deal for them. And even better if they can extract more money beyond those 3 years and then sell what is still left to reclaiming bit more money. Making 15% a year on investment and it fully paying back is not bad move.
Did Broadcom think the VMware moat was a lot deeper than it is? Or do they think losing small or medium customers is worth it to get 10x more revenue from its largest customers?
I'm curious why anyone still uses the product. They had a clear advantage 20 years ago but that was 20 years ago. I really don't understand what distinguishing thing they offer these days.
Not having a clear advantage might be a good reason to not choose it for new products, but there are thousands of companies that chose VMware 20 years ago when it did have a clear advantage. You have to have a pretty good reason to invest thousands of people hours in a project to migrate to another plaform, and simply "it's not the latest product out there" is not often a compelling reason to throw out something that still works pretty well.
Price increases as in the linked article can be a compelling reason, of course.
If a company has a significant amount of infra set up with VMware the costs of migration to another solution are substantial. They have to retrain staff, who may not be keen on relearning how to do what they already know how to do with VMware. They have to replace other solutions that work with VMware but maybe not a new alternative. And there’s usually features that VMware has that others do not, or they work differently enough that it introduces more friction.
The core virtualization stuff is commodity, but there’s much more to it.
A lot of people knows it very well. I wouldn't consider it for personal use but I didn't have a lot of issue managing vsphere clusters when I was still into it.
Also it is not about the VMs only but the whole ecosystem. Last time I worked for a company where we considered moving out of vmware, the main issue was not migrating to a new VM management solution but having to migrate to a new VM hypervisors + migrating to a new backup solution supporting the new hypervisor. This while still maintaining original backup solution until retention runs out or we migrate and test all old backup with long retention. Or we had to decide on maintaining several backup systems in parallel which would increase the cost of backups.
VMware was worth the money because it just worked and worked decently. I chose it in the past to avoid wasting my time on crashes, incompatibilities, and lacking performances that were typical for other VMs.
vmware had this magic sauce that worked before hardware virtualization support was common. In around 2000, the only realistic options was either vmware or vm/esa on a mainframe. It hasn't been like that for a long, long time.
A lot of people use it. I for one use it because it is simple to manage, setting up shared storage is a breeze, live migration is super easy. I don't use any of the fancy NSX stuff and whatnot because I don't believe they add any value.
I was looking at Proxmox, but I kind of got lost in the documentation when it comes to shared storage and high-availability, meaning restarting VMs on another host if one host fails.
RedHat is just as expensive as VMware used to be, but it had less features, so it is not really an option.
If you're talking about OpenShift, it has less features for virtualization perhaps, but is easier for migrating apps over to k8s. Which is something many orgs want to do.
There are two different things:
- live migration: its wonderful because I can update servers firmware without any VMs downtime
- VM restart if a host dies - works wonders in a cluster where given enough host capacity you have a level of assurance that the downtime is very brief.
And regarding Fault Tolerance, it is tricky and if you don't need that level of availability it's better to just plan for a VM restart or have other types of high-availability in place to mitigate a host failure.
Broadcom is the typical MBA run company. Nobody fucking cares about the long run. The idea is to extract as much as much profit as possible right now, the executives pocket the bonuses, and when the well dries, they can just move to the next target.
Broadcom kind of went against this though as it also divested itself of several parts of the vmware product range. A lot of what we used to see advertised alongside sphere isn't even a broadcom product now.
> do they think losing small or medium customers is worth it to get 10x more revenue from its largest customers?
they'd be almost certainly right to think that, of course it'd be worth it. The pending question is whether the largest customers will actually sticks with VMWare for them to actually see this 10x revenue materialise
I would guess they know it's dying and they're trying to get everything they can out of it while they can. Vendor lock-in is a real risk everyone should be cognizant of.
This doesn’t surprise me. The customer in this case is in the hosting business. They are likely fairly expert at infrastructure and devops and are operating in a super competitive, commoditized, price-sensitive corner of the market. Broadcom probably always expected to churn customers like that. But they knew that for every customer like that, there would be 3 or more Fortune-500 type companies that don’t want to risk any disruption and have big IT budgets that they can divert into covering the increase if they have to. They might even make plans to gradually move away over a few year period, but if they are charging 10x more than before, the profit they make in the next couple years will more than make up for it.
Seems to be focused on two different areas. This comparison [0] argues that proxmox is mainly focused on a centralized solution where as OpenNebula is focused on distributed setups
Both seem like excellent options for their use case. At that point, every small benefit helps make a decision. Even the support pricing doesn't seem too far off if you have a reasonably sized cluster.
Both are projects with over a decade and a half of development behind it, both use similar virtualisation architectures, and both have easy to use web interfaces, with distributed storage options and cluster management tools.
I only use it because VirtualBox was being slow and the 3D hardware acceleration wasn't working even when I think version 7 said it should. Switched to VMWare and it's been almost perfect.
Honestly, if I have to choose between working on the codebase of the average enterprise VMWare client and dragging some boxes around in a "no-code platform", I'm taking the boxes.
Broadcom seems to have a lacking understanding of a business of software. Which is a typical symptom of hardware companies, their brains wired differently.
There's no claim about it, it actually happened.
Always amazed at companies that extort their customers because it might be too expensive to move off those products.
I remember Computer Associates back in the 1980s and 1990s, that was their business model, buy out a company, fire all the engineers, and Jack up license fees. Rinse and repeat. I remember seeing them at unix expo at the javits center, they always gave out the best chachkis
We're small, but we abandoned VMWare YEARS ago because (a) our size meant nothing they were doing that was interesting was interesting to us and (b) HyperV was plenty good enough and essentially free and (c) the licensing and support conversation every year even BEFORE Broadcom was just insane.
I get it. Virtualization is a commodity now. They're trying to squeeze the last serious profits out of the platform, but in behaving this way they're absolutely hastening their own exit from the market.
Off topic but if I may. The way people use percentages to express multiples is confusing. A doubling is a 100% increase. So a 200% increase is 3x, and so on. Then at some point we forget about the +1 and 1000% is "10 times the sum it previously paid for software licenses". Just a pet peeve I guess : )
I sat on the unofficial style committee at a company I used to work for. I guess one of our also unofficial tenets was that usages that people regularly used incorrectly or were regularly incorrectly interpreted should be avoided--even if you're "technically" correct in using them. It's the written language equivalent to not using parentheses and depending on people applying, especially less common, precedence operations in code.
So no, not just your pet peeve. I personally hate 100% increase and that sort of thing. Even if I know the correct interpretation, I have no idea if the writer did.
The same kind of people that say "I'm not good at math" will fight to the death to say this the wrong way and not have to think about it, so it's a losing fight, unfortunately.
Now, any time I see anything more than 100% for an increase, I assume the person doesn't know how to say it properly, and the vast majority of the time I'm correct. I certainly never make any decisions (including opinions) based on that information without doing the math myself.
"wrong" and "correct" are only defined by what people actually say - if people mainly say "101% increase" to mean "new value = old value * 1.01" then that's what it means whether it makes sense grammatically or not.
But the point of communication is mutual understanding. I generally agree that language is what we make it be, but in this case the wrong usage, meaning the usage that is mathematically incorrect, is causing issues.
I can't just go around saying that my cat is siamese even if he's not just because I got it wrong one time and now this is what siamese means to me. Because my goal is to transmit accurate messages.
Agreed. Don't express changes in percentages, express them as multiples.
In addition, don't express changes in terms of ratios, express them in terms of actual quantities.
So instead of saying "our efficiency increased by 100%", get rid of the ratio and express it as a multiple of the relevant underlying quantities, e.g. "our costs are 0.5x".
By that argument, 100% is also a factor, but you wouldn't interpret "100% cost increase" as an unchanged cost, so a 10x price shouldn't be described as a "1000% cost increase".
Well, the way I look it is that if someone says 1000% increase, it may be whatever value between x8 and x12, rounded to the round number, to look more dramatic.
Would you really get more value from the headline if it said "923% increase" instead of 1000%?
The article talks about 10 times the price. If that’s true the equivalent percent would be 1100%. So 1000% is a lower, rounded estimate; presumably 10X is also an estimate.
> The article talks about 10 times the price. If that’s true the equivalent percent would be 1100%.
I think you've just proven OPs point. 10x is a 900% increase, or 1,000% of the original price, not 1,100%.
Using percentage for anything other than a fraction of something (i.e. <= 100%) is usually done for effect, often done incorrectly, and even if done "correctly", leads to exactly this kind of confusion.
But % “decrease” is just as bad as percent “increase” that OP is railing against.
Whereas the inverse of 3x is (1/3)x and still far clearer (three times the size and one third the size).
Meanwhile 3x “smaller” is the inverse of 3x “bigger”, not 300% “increase”, which is mathematically different. The adjectives break our brains. Much easier to comprehend if we don’t use them, using “the size” instead.
Is a 200% increase 3x? When it says "200% more" then it's 2x the value on top of the original value, or when it says "levels are now at 200%" then the value doubled. But with "increase" as word, or with random other wordings as in the headline, I'm not actually sure how to take it. I intuitively took it as 10x the original value (not 11x) but I'm now thinking that's probably wrong
Yes. Increase means the number went up. A 200% increase means that it increased by 200% of the original, so you have the original 1x and the increase of 2x for a total of 3x.
“A 200% increase” is logically equivalent to “200% more”. Think about what it means for a number to “increase” or to be “more than” another number. How much is the “increase”? How much “more”? Unfortunately, “200% as much” is often what people mean when they say “200% more”.
(This difference became especially noticeable to me after playing^W practicing my math skills with incremental games for obscene amounts of time.)
It's intentional on the part of the journalist to mislead readers by making the number seem bigger. An off-by-one error is a small price to pay for more clicks.
Would it be so hard to give a ratio? Or the actual numbers? I know people are familiar with percentages, but I think overall the sunny do much for understanding and are really only good as an abbreviation to save repetition, which isn't happening here.
Its confusing, but in this context it is just another way of saying 'huge'. 1000 is a bigger number than 10, so 1000% feels bigger than 10x. The exact numbers don't matter here, so thats why I think people will keep using percentages.
The issue is language ambiguity, not always just math proficiency.
A "price hike" is not necessary an "increase", i.e. a sum of the old price and a change, which is X + X * (P/100). It very well might also mean "a price hike to Y", i.e. "the new price is now P percent of the old" which is a multiple (X * (P/100)).
But yeah, it's usually very confusing for all the parties involved, especially when the change is negative.
A "price hike" is always an increase, and specifically a large increase. If the value has decreased it'd be a "price drop" and a large drop would be a "price slash".
Again, what I meant is, of course, "price hike" in itself always constitutes some increase, but its nature can be one of two: 1. positive delta 2. setting the price to a ratio > 1.
Even worse, people who say ‘3 times less than’ (which would be n - 3n = -2n) when they mean ‘one third as much as.’
And of course, the worst people of all think that one third is less than one fourth.
Personally, I blame the widespread usage of decimal notation rather than fractions. The so-called ‘metric’ system (a misnomer, because every system of measurements … measures) bears a lot of blame here.
We (Red Hat) are doing great business moving customers off VMware. 20,000 VMs is not an unusual amount for our larger customers. I did a 5 min lightning talk in summer about this: https://pretalx.com/devconf-cz-2024/talk/SN93LG/
It is funny because my former company was trying to move out of vmware even before the broadcome aquisition and the local sales rep pretty much told us RHEV was a dead project but at the time openshift virt wasn't ready yet.
In the end I left before broadcom bought vmware but the status at the time was "let's wait a bit more".
I don't agree with the deprecation of RHV since it's very closely modelled on VMware vCenter in terms of the concepts and user interface. So it's ideal for ex-VMware admins. But here we are. Containers & Kubernetes are more buzzword-compliant than virtual machines.
Basically, RHEV as a VMware takeout project never gained a lot of traction. With the general industry shift towards containers and Kubernetes plus the Broadcom acquisition, kubevirt in combination with a much more mature OpenShift is a lot more interesting.
So working in the MSP space providing IT support to clients, our VMs are 90% mostly Windows Servers.
Due to the applications, our clients use + AD + File share. Explain how (excuse the French) the fuck Kubernetes and OpenShift helps here.
Thanks xoxo
Red Hat (obviously) and Microsoft (increasingly) aren't all that focused on Windows as an operating system on the server. And the desktop, in general, is about the browser with some relatively niche exceptions.
Probably, outside of VMware, you run Xen for Windows servers.
Okay, but I thought the whole point of VMs was to be OS agnostic? :)
I should be able to run all OS on a VM platform. :) Right? :) What is the friggin point.
Xen is an option but is it viable for big datacentre deployments?
There are always transition costs from the point of view of migration/management at least. Applies to one version of Linux to another as well.
The strategic jump is probably to Kubernetes/OpenShift/kubevirt but there are probably intermediate steps like Xen one could make instead off VMware.
Odd way of phrasing your question but my general thought is your <10 server, no IT staff MSP customers aren't the focus of this conversation or most of the userbase here.
Looks like that tool typically targets KVM as a source. My understanding is that Red Hat's own virtualization platform is transitioning from RHEV to OpenShift. If that is true, are you actively moving customers from vmware to OpenShift via Kubevirt?
Upstream, virt-v2v supports conversions from VMware to either oVirt (RHV) or KubeVirt (OSV), and we don't plan to drop the oVirt support any time soon.
However Red Hat now only supports conversions to OSV, since RHV was deprecated (sadly).
The biggest problem is oVirt itself has not proven to be a very sustainable open source project. If oVirt dies, we'll likely remove support in v2v. (There's a great start up opportunity here, for a dull but money-making company that productizes oVirt again.)
To move VMs from RHV to OSV you can just copy the disk image since they already should have virtio drivers, qemu guest agent, and be able to boot on any qemu/KVM-based platform. I believe there's some automation for that, but it doesn't involve virt-v2v.
Thank you for taking the time to explain. I was honestly a little shocked to hear about RH moving away from RHV, but glad to hear there is some interoperability between some of these tools.
There never was a lot of interest in RHV as a like-to-like VMware alternative. But, when I left, there was quite a lot of interest (especially post-Broadcom acquisition) in a Kubernetes/cloud-native friendly VM alternative to VMware. And, yes, there are various migration tools though I was not involved in the details.
> There never was a lot of interest in RHV as a like-to-like VMware alternative
That is because it was before Broadcom lit VMware on fire. Nowadays everyone is looking for a like for like VMware alternative.
> Kubernetes/cloud-native friendly VM alternative to VMware
Cloud-native and VMs for the people used to the VMware world are antithetical. Overwhelmingly we're talking about carefully managed pets, which is exactly the opposite of cloud-native.
I agree, hence my idea about resurrecting oVirt (see up thread). Red Hat are all-in on Kubernetes, which is the wrong model for pet VMs, but a better model for gitops stuff.
I'd argue it's still essentially backward facing. Yes. There are buyers who would essentially like a cheaper VMware clone today. But I'm not at all convinced that's a good business model for a company like Red Hat to invest in going forward.
> There are buyers who would essentially like a cheaper VMware clone today. But I'm not at all convinced that's a good business model for a company like Red Hat to invest in going forward.
Yes, they're buyers stuck in the past. If RH can bring them on and take them on a modernisation journey, it could be very lucrative (even if very time consuming).
Thank you for the explanation
Broadcoms policy seems to be to aim for the larger customers where a 10x price rise for a suite of products isn't an issue. Charging 10x the price to 10% of your customers isn't necessarily a bad strategy for a business.
6 months ago on HN: https://news.ycombinator.com/item?id=40452260
> Steve McDowell, chief analyst at NAND research, told The Register that VMware by Broadcom is “laser focused on high-revenue, high-margin business” and has priced its wares “just below the pain threshold for customers they care about.”
There is a famous story about business in the UK when some sales manager at Coca Cola raised the price by a small amount for JD Wetherspoons, one of the largest chains of cheap pubs across the country. The manager must have felt very smug about the extra revenue they'd bring in, from a small change that the client could easily absorb and could hardly moan about. What else could they do, move to Pepsi?
They did exactly that, and the sales manager was fired over losing one of the biggest clients and the loss of revenue that came with it.
One needs to be very, very careful about both the pain threshold and the "** you, I'm moving my business elsewhere" threshold.
Sure, it's a gamble.
But if Broadcom increases price 10-fold and loses half the customers by revenue over the next couple of years, that reflects in their bonus and they can then move on.
The customers might have a plan to move off vmware by 2027, but continue to pay the money until then. Come the 2027 renewal the execs that made the massive bonuses for increasing revenue have left and then when the company says they aren't renewing it's the next generation dealing with the problems, and then they just reach for the three-envelopes.
This really is on the companies that choose a single supplier. If you had two providers, then shifting doesn't take a 3 year program, it's just a policy change to prefer Provider B rather than Provider A.
It's worth bearing in mind that it's not just a financial gamble but a reputational one as well.
Chances are most of that 90% of your customer base who were previously fairly happy are now very unhappy with a forced move to another platform, so in 10 years time when they're in some other business and making purchasing decisions, someone will mention VMWare and the only thing they'll remember is that shit show where they spent 6 months migrating off it because of some ridiculous pricing change.
That’s the big problem I see for them: all of the VMware guys I know _hate_ this move and are talking about being betrayed. Even if they aren’t immediately migrating, I don’t think many people are going to be comfortable increasing their dependency on Broadcom, whereas before that was as close to a given as it gets in business.
Reputational loss won't be realised for a few years, the people making the decisions will have moved on by then.
The problem is that their documentation is now a shit show. Their 10% whales are going to get fed up with that eventually.
When I worked as a SE, we'd often categorized sales reps as 'hunters' vs 'farmers' when it comes to managing the accounts. The hunters would hit their numbers, collect their bonus, move on, and leave the account in shambles for us to try and assist the new rep in recovering. This is the first time I've noticed that attitude in the company leadership. My heart goes out to our (now) Broadcom rep, as I can tell they are putting the squeeze to all contracts, to their mid-term detriment.
For the vast majority of companies, supply chain diversification represents a cost with little immediate benefit. In the long term, it can be a rational hedge against hostile supplier lock-in, but the vast majority of companies aren’t good at making long-term bets.
There really needs to be a different mode of compensation that forces the C-suite to prioritize long-term company success.
Apple[1], for example, has performance based RSU’s for its executives. RSU’s vest over 3-4 years and the number granted is dependent on Apple’s performance relative to the S&P.
Stock prices theoretically have future cash flows built in but markets are inefficient. In this case, if Broadcom starts losing customers in 2027 and revenue declines 5%, it would get hammered in the market at that point in time, punished much more than the 5% drop.
[1] From Apple’s SEC filing:
The value received by our named executive officers in 2022 from long-term equity incentives reflects exceptional stock price performance over the applicable vesting periods. As a result, for the applicable performance periods, performance-based RSUs vested at 200% of target for our named executive officers based on Apple’s TSR relative to other companies in the S&P 500 (“Relative TSR”), which was at: • the 96th percentile for performance-based RSUs held by Ms. Adams, Mr. Maestri, and Mr. Williams, and • the 99th percentile for performance-based RSUs held by Ms. O’Brien.
Mr. Cook was granted an equity award with 75% performance-based vesting and 25% time-based vesting
Theoretically, stock compensation should do that. In practice, many/most investors have a pretty short time horizon in mind as well. I'm not sure how you provide an incentive for long-term company success absent very deferred compensation which I imagine most people reading this wouldn't favor either if it applied to them. Bird in hand worth two in the bush and all that.
And most importantly some clients will not communicate that they are unhappy and just move on quietly. Especially for something like virtualization which is essentially a commodity.
I would guess the plan for the larger companies is to sell a whole suite of products around the VMware: management, ops, auditing etc, plus take very good care of their niche requirements so moving becomes a much more expensive proposition
Oh interesting.
In sweden it's really hard to find a pub that has coca cola, they almost exclusively have pepsi.
Also in the supermarket pepsi is cheaper.
Wetherspoons might well be the sort of company Broadcom wouldn't want as a customer.
But does it not make you more vulnerable to competitors? Losing one of your few big customers will hurt regardless, but it likely hurts less if you have a pool of SMBs to act as a buffer.
In the long term, sure. They might just be strip mining it.
In terms of competition it's really the same. Whether you fight with a team of 1 for 5 businesses or a team of 10 for 1 business that is 50x bigger, on average as long as you have equal competition in each segment you walk away with the same. But you get very different customers, for better or worse.
A segment of a customer base is hardly a buffer, if you underserve them they will leave, and you will have wasted effort making an inclusive sales and support process to them.
That makes not losing those customers more critical: if you lose one of the 50 smaller businesses, your revenue is barely affected; if you lose that 50x whale, you’re out of a job especially given how hard it is to find another one of that size.
This shows up especially in the context of hiring: for a generation, there have been a ton of people who learn VMware at businesses of all sizes and your large shop can reliably hire them. Shutting down smaller sales affects that directly and because the smaller places are going to do something else it means that the people who move up to a bigger place are going to have direct firsthand experience contradicting what the VMware sales people are saying about how risky it is not to use their software.
This doesn't sound exactly like a small customer, with 20000 virtual servers on 3000 physical ones isn't exactly a tiny deployment.
It’s definitely going to be on the small side.
My first reaction was that you're wrong, but you're probably correct in that large scale VMware installations are much larger than 3000 servers. The small side is going to be a large set of customers, probably the majority.
There are tons of VMware installation that are 3 - 20 physical server, more so than 10.000 server installations. Right now Broadcom is telling those customers, and those with 500, or 1000 servers, that they can take their business elsewhere. Broadcom is much happier to milk the top 5% of customers, compared to servicing the "bottom" 95%.
Broadcom seems to be cutting of to large a chunk of customers. They also are preventing new large customers from coming in. No one starts out with 10.000 physical servers and no one wants to switch a growing 5000 server datacenter from HyperV, KVM or Nutanix to VMware.
In this case you're probably better off thinking of these things in roughly logarithmic terms rather than absolute. By the time you're up to 20,000 servers, you are going to have roughly the same issues to solve as a 10,000 server install and a 30,000 server install. In absolute terms there may be VMWare installations that seem much bigger, but in logarithmic terms there aren't that many more classes above them, and they get rarified quickly. The same ways that you're telling this customer to get stuffed are going to be affecting even the bigger guys in a way that you can't necessarily compare the experience of this customer to the "10 VMs on one server" customers. It's risky to do that. I'm not sure there's as much differentiation in the market to be exploited as Broadcom thinks. Even their larger customers are not just going to sit there and take it... they're just going to move more slowly. But at that scale, VM migrations have a very natural granularity to them; you move them one at a time. You don't need to drop VMWare next quarter; you can just begin your migration project and move away at a pace of your choosing.
But then again, Broadcom may well be explicitly viewing this internally as "we no longer have a competitive product, there is either no way to differentiate it again or we're not willing or able to invest in what it would take, and so the game is just to milk all the money we can out of this", in which case this is pretty close to what I'd expect to see. Where I personally sit in their customer base now, VMs are a commodity and I've got no reason to pay for them anymore. That's been true for a while. I'm speaking here just as A Guy, not as an employee of anyone, but I would expect that this reality has been slowly but surely crawling up the market from where I sit at the bottom.
(This is just discussion, not "disagreement".)
Hell, AT&T just settled with Broadcom. Five bucks says they stay right where they are. Hock Tan knows what he’s doing to force the big players to pay and extort them to the nines.
https://www.ciodive.com/news/broadcom-att-vmware-settlement-...
A less charitable (but I'd argue, likely more correct on balance of probabilities) is they're aiming for the larger customers who are so entrenched in their use of the VMWare product stack that they can't feasibly migrate without taking a huge one-time hit to their balance sheets.
i never understood, even 15 years ago, why people still use vmware when so many open source (i.e. virtualbox) systems existed at the time. The fact that Broadcom buys it out and then 10x the price on people is just even more reason to move to open source. As noted in the article, the open source solutions are 2x faster anyway, so they customers can realize a 2x increase in slices per server.
The ones staying with VMware, permanently or near term, are likely using products for which there are not good replacements. The open source alternatives can do what ESXi does sure, but its progressively more difficult to replicate VSAN (including stretched clusters) and NSX-T (with distributed firewalling) without bringing in multiple vendors that will all point at each other when there is a problem.
Vmware is better by our performance tests. Still is better but not 10x the price better.
There are some open source options that come close to what vmware's offerings could do, but virtualbox for sure ain't that. It has an extremely limited feature set compared to other offerings in this space. For example, it is mainly intended to manage VMs on a single host. Most organizations looking at commercial offerings have multiple hosts (or racks, or even datacenters) they want to manage. Apples are much more similar to oranges than vmware is to virtualbox.
virtualbox is buggy and it's not open source.
The VirtualBox extension is not open source, but VirtualBox is GPLv3: https://www.virtualbox.org/svn/vbox/trunk/COPYING
I can't speak to your buggy claim since I don't know which feature you are complaining about, and if it's part of the extensions then the "many eyes, shallow bugs" does not apply
Because they dumb it down, you pay more for VMware but then you can hire cheaper, less educated people to run it.
For desktop virtualization, vmware workstation is the best. Full stop.
[flagged]
VMWare Workstation is now free, and VMWare Player has been discontinued. I'm not sure yet what the catch may be.
The desktop is mostly a pretty uninteresting space in general. Virtual desktops never really took off (which VMware was into at one point) and desktop virtualization is generally a commodity for the relatively few people who use it.
not smart if you make it hard to use and learn and the next generation of technologists gain experience and familiarity with other solutions and improve them to be superior products.
Bold of you to assume Broadcom cares about long term sustainability.
From what I have seen they are just a PE firm pretending to be a tech company. They will bleed the customers who can’t move fast enough. Burn it to the ground to get every last dollar.
Unless they have a monopoly on the tech, that's not how the free market works. Raising your prices x10 without having a government mandated monopoly, is a great incentive to create competitors.
And so what? Companies coming in to fill the gap of the market you're leaving isn't a concern
They then make their product good and convice your customers they are good enough for less. They don't eken have to be better, just go enough.
So many people misunderstand what is actually happening.
Broadcom aren't hiking the price 10x for everyone.
If you previously used:
Then you're going to see you vSphere costs jump massively because you're only option to buy vSphere now also comes with: So, if you're already heavily invested in VMWare, you might even see your costs reduce.From Broadcom's perspective, that seems to make even less sense than a simple price hike that lets them shed unprofitable customers. What's the point of doing that? Some kind of sales force optimization?
Everyone now has to ask the question, should i continue paying for product that does XYZ from VMWare's competitor when i'm already getting a product that does XYZ as part of the VMWare licence.
Chances are if you're a vmware estate, those products are going to integrate better and then you only have a supplier, support portal, customer rep, etc..
Broadcom loose customers who arn't really invested in to VMWare and those who stay become even more invested in to them.
>those who stay become even more invested in to them
Maybe. Or maybe they understand the play being made as well as everyone else does, and won't be inclined to become any more dependent on an organization that will turn the screws nice and hard on anyone they know won't be able to leave them.
> What's the point of doing that
A sale isn't free. It requires time, headcount, and engineering+PM resources.
At Broadcom size, you want to maximize sales to F1000 customers (the kind that can and will shell 7-8 figures in TCV) as that gives you the most bang for your buck from a SalesOps perspective.
A major issue with VMWare was their sales team would chase logos at the expense of an optimal sale, so their account teams would give extreme discounts (70-80%) at barely above the margin in order to close deals.
Well, we're probably a F10 company and my project had to move off of our vmware environment because prices were too high. Even at enormous companies, these things matter, and there's still accounting for resources at the local level.
I work at an IT outsourcing company that also has some VMWare farms (though not sure if we reach 20k VMs).
What I picked up is that before the acquisition, price was mostly determined by CPUs on the hypervisor, now it's mostly by RAM.
Which means we now have to optimize hypervisor hardware to a different licensing model.
I don't think we've seen 10x price hikes, but it could possible if a provider has optimized hard for a metric that is now less relevant, and/or has difference licensing deals.
Caveat: I've picked all that up second-hand through the grapevine, I'm not working with our virtualization farms directly. So not 100% on top of what's going on.
It's the Oracle pricing model. Turn the customer upside down and keep shaking until you've got all their money. They will dress this up in different ways to different customers, but that's the model.
> price was mostly determined by CPUs on the hypervisor, now it's mostly by RAM.
I think you might have that backwards, before we paid based on ram and now we pay based on cores
> and now we pay based on cores
Didn't even oracle drop this monstrosity of a licensing fee? Didn't microsoft also dropped / stoped pursuing it? How on earth do people justify this?
Having been in the field for some time and having some contacts, I must say: 100% of the companies where I get some insights are currently either already building alternatives or reviewing their options with a strong intention to leave VMware behind. And in many cases this is quite a huge task, but certainly doable with the required momentum.
I'm in small Devops team and people are pissed off. Thankfully Proxmox works just as well for us (some would say better). Even a large-scale migration seems scriptable to me.
Proxmox looks great, but unfortunately it doesn’t come with encryption oob and manually applying it comes with feature loss and other problems. ZFS vs encryption is a bad place to live. You of course need encryption for your VMs on the Enterprise.
Proxmox is a lovely piece of software, highly recommend to everyone. But I think Proxmox and vSphere target rather different scales of cluster. Proxmox starts to choke once your past 20 nodes or so. Corosync just wasn't designed for huge clusters.
I think the price hikes have made the "medium" use-case more uncertain.
I know VMWare works great for a lot of people, but it was always just a thorn in my side. Almost a decade ago my small company moved from VMWare to ganeti, and we've been really happy with it.
At the time we were running a SAN and the small VMWare Essentials license. But managing it with Windows was painful for me personally, and our HP SAN was kind of painful, every year like clockwork it would just stop processing until we had remote hands reboot it. And every year they'd tell us "Try this firmware" to shut us up for another year, until eventually the product reached EOL.
We had been running ganeti in our dev/stg cluster for years at that point (and the python.org infrastructure was using it at that time as well), so I decided to transition us to local storage running VMWare and then to ganeti. Doing local RAID array monitoring on VMWare was quite painful at the time, maybe that's gotten better?
I absolutely love that ganeti has such a good suite of command-line tools for operating. I have a small shell script that respins half our dev/stg infrastructure every other day. Our entire critical infrastructure is HA at the machine level, so we don't play around with anything like live VM moves, we just do offline moves. I used to do the drbd-based replication and live moves, but the doubling of disk writes and committed disc usage wasn't worth it for us.
Proxmox is on my radar if we needed to move away from ganeti, but so far we are looking like we'll stick with ganeti.
Is ganeti still alive? I remember it being a thing a decade+ ago, but haven't seen it mentioned much since.
The latest release is almost 3 years ago, but the repo has commits a month ago. So I guess it depends on your definition of "alive". As "stable" software, I'm fine with it not having a ton of active development. It's been a real workhorse for me, I've got it running on 2 clusters of 6 physical machines each (1.4TB total RAM, 50TB disc, each).
~6 years ago I went to a local Dell dog and pony show where someone at the next table was poo-pooing on "those greybeards that run Debian and ganeti" instead of VMWare. But, this shit works...
Good to hear. VMware used to be honestly my preferred option because it handled certain hairy things well out of the box that many other solutions either ignore or require non trivial integration work (for example it's trivial to connect a VMware VM to piss-poor external network by simply attaching a VLAN - which matters when the network is a fieldbus in some OT deployment)
Ganeti I remember as comparatively harder to deploy (nothing was harder than openstack though)
So Broadcom is basically just milking VMware until it completely dies? I guess you should always consider the risk of a company buying a product you use and then doing this
VMWare has been dying a death of 1000 cuts for years. It's been passed around to now the third owner. They created an entire market and have been watching it erode away.
VMware has been increasing revenue year after year [1]. That seems quite the opposite of dying?
[1] https://www.netcials.com/financial-revenue-history-usa/11246...
This is exactly the problem with optimizing metrics though. It seems like they’re getting more money from less customers but it puts them at might higher risk that those big customers will churn and tank the company which also increases as the price goes up as it will eventually hit a point where they can find something cheaper.
> VMware has been increasing revenue year after year [1]. That seems quite the opposite of dying?
And what have its profits been over the same time period?
Also, how did maximizing numbers work out for Boeing?
not if it keeps losing market share.
A rising tide makes the pie bigger. Something like that.
Is virtualization really a growing pie, these days? Big players eat everything, and they have their own solutions typically.
The other possibility, if it's a shrinking market, is they are just strip-mining it.
VMware are the big players
I think he means AWS/Azure/GCP and other hyperscalers.
That's what enshittification looks like right before the company dies.
Yeah but how does it make any business sense for Broadcom to be doing this? They could have raised prices 100% and fewer people would flee and they'd make more money overall. 1,000% price raises are just plain extortion.
If the action of a company doesn't seem to make sense, ask yourself: could you imagine an incentive structure within the company that could lead to such actions?
Buy, extract as much money as possible, then dump the carcass. I'm sure they did the math on this.
Whenever companies make moves that people don't like, the comments go on as if they haven't thought any of it through
They knew they would lose customers, they've factored that into their business plan
Considering half the comments are “Proxmox” and “Virtualbox/Parallels” are so much better anyway, I’m guessing most commenters in here are extremely naive and not aware of the IT landscape.
This is the classic "golden goose" strategy. Why wait around all week for steady returns when you can just rip the thing open and see what you can get for the internal organs.
Mathematically if they can extract say 1.5 times the money they used to buy something in 3 years. It is a good deal for them. And even better if they can extract more money beyond those 3 years and then sell what is still left to reclaiming bit more money. Making 15% a year on investment and it fully paying back is not bad move.
Raise prices. Layoff staff. Less support. = more cash flow (Short term).
In order to sell?, spinoff?, or use the cash to finance something else?
Did Broadcom think the VMware moat was a lot deeper than it is? Or do they think losing small or medium customers is worth it to get 10x more revenue from its largest customers?
I'm curious why anyone still uses the product. They had a clear advantage 20 years ago but that was 20 years ago. I really don't understand what distinguishing thing they offer these days.
Not having a clear advantage might be a good reason to not choose it for new products, but there are thousands of companies that chose VMware 20 years ago when it did have a clear advantage. You have to have a pretty good reason to invest thousands of people hours in a project to migrate to another plaform, and simply "it's not the latest product out there" is not often a compelling reason to throw out something that still works pretty well.
Price increases as in the linked article can be a compelling reason, of course.
If a company has a significant amount of infra set up with VMware the costs of migration to another solution are substantial. They have to retrain staff, who may not be keen on relearning how to do what they already know how to do with VMware. They have to replace other solutions that work with VMware but maybe not a new alternative. And there’s usually features that VMware has that others do not, or they work differently enough that it introduces more friction.
The core virtualization stuff is commodity, but there’s much more to it.
A lot of people knows it very well. I wouldn't consider it for personal use but I didn't have a lot of issue managing vsphere clusters when I was still into it.
Also it is not about the VMs only but the whole ecosystem. Last time I worked for a company where we considered moving out of vmware, the main issue was not migrating to a new VM management solution but having to migrate to a new VM hypervisors + migrating to a new backup solution supporting the new hypervisor. This while still maintaining original backup solution until retention runs out or we migrate and test all old backup with long retention. Or we had to decide on maintaining several backup systems in parallel which would increase the cost of backups.
It is not just about migrating VMs.
VMware was worth the money because it just worked and worked decently. I chose it in the past to avoid wasting my time on crashes, incompatibilities, and lacking performances that were typical for other VMs.
Nobody got fired for buying IBM/Microsoft/Cisco/VMWare
Inertia, both in a company and in the industry, is a major reason.
Maybe simply the moat? Integration with networks? What is the alternative? Xen? Proxmox?
Depends on what you're doing I guess. https://en.wikipedia.org/wiki/Comparison_of_platform_virtual... there's a bunch
vmware had this magic sauce that worked before hardware virtualization support was common. In around 2000, the only realistic options was either vmware or vm/esa on a mainframe. It hasn't been like that for a long, long time.
>In around 2000, the only realistic options was either vmware
Let's not forget user mode linux. :-) I ran hundreds of "VMs" using UML and it worked surprisingly well.
A lot of people use it. I for one use it because it is simple to manage, setting up shared storage is a breeze, live migration is super easy. I don't use any of the fancy NSX stuff and whatnot because I don't believe they add any value.
I was looking at Proxmox, but I kind of got lost in the documentation when it comes to shared storage and high-availability, meaning restarting VMs on another host if one host fails.
RedHat is just as expensive as VMware used to be, but it had less features, so it is not really an option.
If you're talking about OpenShift, it has less features for virtualization perhaps, but is easier for migrating apps over to k8s. Which is something many orgs want to do.
> ..., meaning restarting VMs on another host if one host fails.
Wait, your VM will still restart if vmware host dies, right? You can't have RAM sync between hosts.
There are two different things: - live migration: its wonderful because I can update servers firmware without any VMs downtime - VM restart if a host dies - works wonders in a cluster where given enough host capacity you have a level of assurance that the downtime is very brief.
And regarding Fault Tolerance, it is tricky and if you don't need that level of availability it's better to just plan for a VM restart or have other types of high-availability in place to mitigate a host failure.
> You can't have RAM sync between hosts.
That's what FT is for, if the primary VM instance dies or becomes unresponsive, you failover to the mirror instance.
Broadcom is the typical MBA run company. Nobody fucking cares about the long run. The idea is to extract as much as much profit as possible right now, the executives pocket the bonuses, and when the well dries, they can just move to the next target.
It is the locust economy.
Broadcom aren't charging 10x for the same thing.
They are just not selling products individually anymore.
If you previously only licensed a single product, your only option now is to buy everything under a single license.
We are already using a lot of VMWare products, not only do we now have access to more, our overall price went down.
I think medium sized customers have a choice, switch from their Broadcom products, or move more to Broadcom.
Broadcom kind of went against this though as it also divested itself of several parts of the vmware product range. A lot of what we used to see advertised alongside sphere isn't even a broadcom product now.
> do they think losing small or medium customers is worth it to get 10x more revenue from its largest customers?
they'd be almost certainly right to think that, of course it'd be worth it. The pending question is whether the largest customers will actually sticks with VMWare for them to actually see this 10x revenue materialise
The latter. They have a threshold where they basically say IDNGAF. Above that, they’ll come to some terms.
I would guess they know it's dying and they're trying to get everything they can out of it while they can. Vendor lock-in is a real risk everyone should be cognizant of.
Is https://www.theregister.com/2024/12/02/beeks_group_vmware_op... a better link to use? This article seems to just be a summary of it.
This doesn’t surprise me. The customer in this case is in the hosting business. They are likely fairly expert at infrastructure and devops and are operating in a super competitive, commoditized, price-sensitive corner of the market. Broadcom probably always expected to churn customers like that. But they knew that for every customer like that, there would be 3 or more Fortune-500 type companies that don’t want to risk any disruption and have big IT budgets that they can divert into covering the increase if they have to. They might even make plans to gradually move away over a few year period, but if they are charging 10x more than before, the profit they make in the next couple years will more than make up for it.
Why not proxmox instead? I have never heard of OpenNebula before
Seems to be focused on two different areas. This comparison [0] argues that proxmox is mainly focused on a centralized solution where as OpenNebula is focused on distributed setups
[0] https://cylab.be/publications/64/download/2024-comparison-of...
So, their whole decision is based off of which VMM can support qcow import out of the box
Both seem like excellent options for their use case. At that point, every small benefit helps make a decision. Even the support pricing doesn't seem too far off if you have a reasonably sized cluster.
Both are projects with over a decade and a half of development behind it, both use similar virtualisation architectures, and both have easy to use web interfaces, with distributed storage options and cluster management tools.
I fished this out of the cylab PDF posted below: https://github.com/OpenNebula/one#readme (Apache 2)
> Why not proxmox instead? I have never heard of OpenNebula before
Different type of architecture. XCP-ng would also be a closer to fit to how VMware works as well.
(I currently run Proxmox at work. At my last job we used OpenStack.)
Why go linux? MNX Triton and vanilla SmartOS are good hypervisor alternatives too.
Or xcp-ng + xen orchestra.
xcp-ng is far from stable and lacks significant features
I only use it because VirtualBox was being slow and the 3D hardware acceleration wasn't working even when I think version 7 said it should. Switched to VMWare and it's been almost perfect.
Sounds like Broadcom is engaging in some serious enshittification of VMWare. Customers leaving in droves is the appropriate reaction.
Perfect opportunity to start transitioning from 20,000 VMs to the actual amount needed for most reasonable use cases - 0.
Sounds like someone wants to go full no-code to me.
Honestly, if I have to choose between working on the codebase of the average enterprise VMWare client and dragging some boxes around in a "no-code platform", I'm taking the boxes.
Seen this dozens of times this year. This is policy from VMware/Broadcom.
Broadcom seems to have a lacking understanding of a business of software. Which is a typical symptom of hardware companies, their brains wired differently.
Milking people for millions is poor understanding?
There's no claim about it, it actually happened. Always amazed at companies that extort their customers because it might be too expensive to move off those products. I remember Computer Associates back in the 1980s and 1990s, that was their business model, buy out a company, fire all the engineers, and Jack up license fees. Rinse and repeat. I remember seeing them at unix expo at the javits center, they always gave out the best chachkis
We're small, but we abandoned VMWare YEARS ago because (a) our size meant nothing they were doing that was interesting was interesting to us and (b) HyperV was plenty good enough and essentially free and (c) the licensing and support conversation every year even BEFORE Broadcom was just insane.
I get it. Virtualization is a commodity now. They're trying to squeeze the last serious profits out of the platform, but in behaving this way they're absolutely hastening their own exit from the market.
Good riddance.
only one?
1k% = 1 kilo% = 1 000% = 11? times price hike.
1k% = 1 kilo% = 1e-2 kilo = 3.22 troy ounces
only one company?
Off topic but if I may. The way people use percentages to express multiples is confusing. A doubling is a 100% increase. So a 200% increase is 3x, and so on. Then at some point we forget about the +1 and 1000% is "10 times the sum it previously paid for software licenses". Just a pet peeve I guess : )
I sat on the unofficial style committee at a company I used to work for. I guess one of our also unofficial tenets was that usages that people regularly used incorrectly or were regularly incorrectly interpreted should be avoided--even if you're "technically" correct in using them. It's the written language equivalent to not using parentheses and depending on people applying, especially less common, precedence operations in code.
So no, not just your pet peeve. I personally hate 100% increase and that sort of thing. Even if I know the correct interpretation, I have no idea if the writer did.
Well "100% increase" is unlikely to have been written by a mistaken writer. It's any number above 100% where it's possible the writer made an error.
And doubled would be even simpler. Yes 100% is probably mostly unambiguous but I'm certainly unsure about 131%.
> usages that people regularly used incorrectly or were regularly incorrectly interpreted should be avoided
Lordy, YES. A successful business must have good communication. Not people saying "actually, I was right, but ...".
The same kind of people that say "I'm not good at math" will fight to the death to say this the wrong way and not have to think about it, so it's a losing fight, unfortunately.
Now, any time I see anything more than 100% for an increase, I assume the person doesn't know how to say it properly, and the vast majority of the time I'm correct. I certainly never make any decisions (including opinions) based on that information without doing the math myself.
> the vast majority of the time
"wrong" and "correct" are only defined by what people actually say - if people mainly say "101% increase" to mean "new value = old value * 1.01" then that's what it means whether it makes sense grammatically or not.
But the point of communication is mutual understanding. I generally agree that language is what we make it be, but in this case the wrong usage, meaning the usage that is mathematically incorrect, is causing issues.
I can't just go around saying that my cat is siamese even if he's not just because I got it wrong one time and now this is what siamese means to me. Because my goal is to transmit accurate messages.
Caveat emptor may be legal in capitalism, but it is not a great way to be a coworker which was the point.
Agreed. Don't express changes in percentages, express them as multiples.
In addition, don't express changes in terms of ratios, express them in terms of actual quantities.
So instead of saying "our efficiency increased by 100%", get rid of the ratio and express it as a multiple of the relevant underlying quantities, e.g. "our costs are 0.5x".
I am a 100% programmer
Open source, fully documented, and maintainable?
Not my domain; thats for the legal department, the documentation department and the maintenenace team
> don't express changes in terms of ratios, express them in terms of actual quantities.
Would be lovely, but transparency is not good for every way of doing business :’ )
That's more a language issue than a number issue. 1000% is already a factor, it's literally the number 10.
The problem is that the whole sentence doesn't make sense with a number there. "Company claims 10 price hike drove it from VMware..."
By that argument, 100% is also a factor, but you wouldn't interpret "100% cost increase" as an unchanged cost, so a 10x price shouldn't be described as a "1000% cost increase".
“10x price hike” would have been totally clear. “Factor of 10 price hike” would probably be a formal way to say it.
Well, the way I look it is that if someone says 1000% increase, it may be whatever value between x8 and x12, rounded to the round number, to look more dramatic.
Would you really get more value from the headline if it said "923% increase" instead of 1000%?
Over a 9x increase or almost a 10x increase would be just as dramatic and less mental effort and uncertainty to decode.
The article talks about 10 times the price. If that’s true the equivalent percent would be 1100%. So 1000% is a lower, rounded estimate; presumably 10X is also an estimate.
> The article talks about 10 times the price. If that’s true the equivalent percent would be 1100%.
I think you've just proven OPs point. 10x is a 900% increase, or 1,000% of the original price, not 1,100%.
Using percentage for anything other than a fraction of something (i.e. <= 100%) is usually done for effect, often done incorrectly, and even if done "correctly", leads to exactly this kind of confusion.
Just posted this exact thing and deleted it. Your’s is better.
My pet peeve goes in the opposite direction: the use of multiples and percentages greater than one to describe smaller numbers: "This is 3x smaller."
But % “decrease” is just as bad as percent “increase” that OP is railing against.
Whereas the inverse of 3x is (1/3)x and still far clearer (three times the size and one third the size).
Meanwhile 3x “smaller” is the inverse of 3x “bigger”, not 300% “increase”, which is mathematically different. The adjectives break our brains. Much easier to comprehend if we don’t use them, using “the size” instead.
333.33% this - drives me nuts.
it's also because a 200% increase is 3x, but 200% of the old price is 2x
Is a 200% increase 3x? When it says "200% more" then it's 2x the value on top of the original value, or when it says "levels are now at 200%" then the value doubled. But with "increase" as word, or with random other wordings as in the headline, I'm not actually sure how to take it. I intuitively took it as 10x the original value (not 11x) but I'm now thinking that's probably wrong
> Is a 200% increase 3x?
Yes. Increase means the number went up. A 200% increase means that it increased by 200% of the original, so you have the original 1x and the increase of 2x for a total of 3x.
What's a 0% increase? The original quantity. Therefore a 200% increase has to be 3x.
“A 200% increase” is logically equivalent to “200% more”. Think about what it means for a number to “increase” or to be “more than” another number. How much is the “increase”? How much “more”? Unfortunately, “200% as much” is often what people mean when they say “200% more”.
(This difference became especially noticeable to me after playing^W practicing my math skills with incremental games for obscene amounts of time.)
30% means 0.3. A 30% increase is x * 1.3. Why should it be different for 200%
Because it's relative to the original value.
Your logic: 100% means 1. A 100% increase is x * 1.
Correct logic: 100% means 1. A 100% increase is x + (x * 1)
Your 30% example is the same x + (x * 0.3)
200% is x + (x * 2) for a total of 3 times the original amount.
Same goes for 1000% being x + (x * 10) for a total of 11 times the original amount.
Well it depends if you mean increase or not. in the OP it could mean either („X% price hike“ could be interpreted as new price is X% of old price)
Switching to a multiple doesn't change the confusion either. A 10x increase, is x + 10x in the same way a 1K% increase would be x + 10x.
"1000%" isn't much more than 900% in the grand scheme of things, it's just 10% more - and that's before likely rounding errors it conveys the scale.
"200%" increase vs "100%" increase is double the difference, a far more significant number.
It's intentional on the part of the journalist to mislead readers by making the number seem bigger. An off-by-one error is a small price to pay for more clicks.
Would it be so hard to give a ratio? Or the actual numbers? I know people are familiar with percentages, but I think overall the sunny do much for understanding and are really only good as an abbreviation to save repetition, which isn't happening here.
Its confusing, but in this context it is just another way of saying 'huge'. 1000 is a bigger number than 10, so 1000% feels bigger than 10x. The exact numbers don't matter here, so thats why I think people will keep using percentages.
In between that and people not understanding the difference between % and pp, I avoid using percentages at all unless I have to.
The issue is language ambiguity, not always just math proficiency.
A "price hike" is not necessary an "increase", i.e. a sum of the old price and a change, which is X + X * (P/100). It very well might also mean "a price hike to Y", i.e. "the new price is now P percent of the old" which is a multiple (X * (P/100)).
But yeah, it's usually very confusing for all the parties involved, especially when the change is negative.
> A "price hike" is not necessary an "increase"
A "price hike" is always an increase, and specifically a large increase. If the value has decreased it'd be a "price drop" and a large drop would be a "price slash".
Yeah, that's the point, ambiguity.
Again, what I meant is, of course, "price hike" in itself always constitutes some increase, but its nature can be one of two: 1. positive delta 2. setting the price to a ratio > 1.
They do this because it makes it look like an even bigger increase.
Yeah, but a 1000% increase is roughly a 10x increase. Good enough for me.
Even worse, people who say ‘3 times less than’ (which would be n - 3n = -2n) when they mean ‘one third as much as.’
And of course, the worst people of all think that one third is less than one fourth.
Personally, I blame the widespread usage of decimal notation rather than fractions. The so-called ‘metric’ system (a misnomer, because every system of measurements … measures) bears a lot of blame here.
[flagged]