Research suggests more than half of VMware customers are looking to move Research published by Civo indicates that more than half of VMware customers are considering leaving the platform under Broadcom's ownership. The VMware rival has an ax to grind here. CEO Mark Boost told El Reg earlier this year that Broadcom's strategy appeared to consist of ditching smaller accounts via price hikes while …
I never trust these "xx% of people asked said they were considering ..." type surveys. If you'd asked me any time during the 20 years I worked at my last full-time job whether I was considering leaving the answer would always have been "Yes". There's a gulf of difference between considering something and actually doing it.
You are correct, however, I am currently consulting with a very large pan-European client who is actively looking at alternatives, given their VMware bills are coming in at something like 8-10x what they were.
This is much more than a maybe. There is nothing in VMware that excludes them from other offerings (they don't leverage any of their expensive components). Of course that may be trickier for some consumers of VMware.
I'm both in agreement with you that it should be easy these days, but that it's also terribly hard because you need to get your techs used to something new. And sometimes us techies stick with the old because it's comfortable and we know it. And we're terrified of being made obsolete by new things. I've worked for years with people who insisted (resisted?) on using the GUI to do long tasks because they didn't want to learn the CLI or scripting to help automate their lives. Because it was new or different or strange or just _harder_ than mindlessly clicking on GUI buttons.
And when you're clicking away madly, you're getting work done! And you can show people! And you can look busy!
Me, not so much when I automate crap and it just happens and I get to surf the web looking for new techniques to automate my life.
This. With a GUI if something breaks, you can always uncheck the box you checked that broke it...you don't have to understand anything and you can log a ticket against the support contract and let someone else debug it with the logs you dumped and emailed to them.
It's sad but to make up the numbers in the industry we've had to flood the decks with dross.
It's worse in software development, the sheer number of developers I've worked with that fight Windows to write Python scripts using Cygwin is unbelievable...and then they push a release that is only supported on Windows...yet it's Python and can run anywhere.
It really makes me want to curl up in a ball in a corner somewhere and cry sometimes. I have clients that make me feel so very alone in the universe.
They're all in the same club as the folks that won't use anything other than VSCode because they can't function without autocomplete.
VSCode isn't the best IDE out there though...I like my IDE to comfy, like a welcoming couch...VSCode is like a board room chair...it's alright when you first sit down, but after a while you realise the chair isn't there for comfort, it's there because it looks good in a board room.
What matters to me in an IDE is solid syntax highlighting, crispy font rendering, smooth scrolling, a terminal pane and a project tree with different colours for different file formats and the ability to customise the appearance to suit my eyeballs...that's it...no distractions, no superfluous navigation bars, no "quick run" buttons...
Less is always more with an IDE...I don't need an IDE that is like the cockpit of a fighter jet, I need an IDE that is like the drivers seat of a Bentley...I just don't want or need millions of buttons and features to fiddle with to make me look & feel busy...I don't want to feel busy and definitely don't want to actually be busy...I just want to write code and run it.
I've been building things for 20+ years now and I've got a whole suite of my own tools for automation and debugging and such that are built just for me...I'm a programmer and a techie...people pay me to solve problems and build them things...and they pay handsomely...if other people are willing to pay a premium to hire me, it should go without saying that I'm willing to hire myself to solve problems and build things too...because of this, I have hundreds of thousands of pounds worth of custom tools and scripts that wouldn't be anywhere near as effective if I'd bought stuff off the shelf that isn't tailored to me.
If I do find something "off the peg" that is interesting, I'll reverse engineer it and build my own version of it for two reasons. Firstly, I want to know how it works because knowing how things work is probably 90% of my job...secondly, I want to tailor it to my own needs.
I'll give you an example of something I built based on something that exists because I thought the existing product was great, but a tailored version works much better. Push Bullet. Great tool, great idea...originally it was a very simple tool...it was essentially an API to which you could "push shit" and have feeds of automated data that you could push an app on a phone...love the idea...but then it became a bridge for your SMS messages, a chat client, a social network...it became something corporate and over the top...you can still "push shit" to it and use it for feeds of automated data...but all the extra crap that exists now makes it a bit of a clusterfuck and a lot less elegant...so I built my own...the way I tailored it for my purposes, I added the ability to specify the kind of data I'm pushing and added the ability to graph things, which I can specify in a flag on the API...so on my phone app (also custom built), I can pick a feed from a drop down and it immediately displays the way I expect it to....if I push time series data to it, I can set a flag that shows it as a graph (line, bar etc), if it is debug / telemetry information from a service I have running somewhere, I can flag it to display as a timestamped log etc etc...so it has become a multifunction tool of sorts...I can use it to track software running in production, I can follow the Bitcoin price in a line graph, I can view customer backup logs as a timestamped log, I can track AWS usage in a bar chart...whatever I decide...and all I have to do is push it to the API with a simple script triggered by cron and select it from a drop down on my phone and boom it displays...I can also write simple scripts that track specific errors in server logs and graph their frequency to find patterns etc etc...I use it with my own custom network auditing tool (which is built from a Radxa 3E), I can plug it in, it will run my custom multithreaded network scanning script...grab banners, check open ports etc etc...which gets winged off to my API, flagged a "network report", it gets spun through an LLM to summarise everything, then I get a nice PDF I can access from my phone app...so if I visit a prospective client, I can have their network audited in under 10 minutes (with reasonable detail, I'll be honest, it's not amazingly detailed, but it's enough to have a discussion on the spot and answer questions that I would have asked the client)...the key thing with this is that I have information *before* I sit down with them rather than having unanswered questions, which I will answer with an audit *after* the meeting followed by a second meeting...it cuts down on wasted time...
It's kind of halfway between Pulseway and Pushbullet with none of the bollocks and 100% privacy.
I recently added the ability to pipe data into Ollama to generate summaries.
If VMWare provided no value, businesses wouldn't use it. And downplaying the complexity and risks of migration serves no one: budget to test and plan.
The free stuff is only good if you have the resources yourself to maintain and support it.
Cheaper, yes. Open source, fine. But businesses will also want support and will be happy to pay fairly for it.
"There's a gulf of difference between considering something and actually doing it"...ironically that is also the difference between shit pay and good pay in this industry.
Techies are notoriously risk averse. Right from the beginning of their careers when they ask questions like "which is the best programming language to learn?" or "should I get a CCNA?"...the anxiety levels amongst tech professionals is real...and the older they get, the less risk averse they become.
I've worked with tons of engineers and techies at various points in their careers...and it is extremely rare to find one that is willing to just walk away from something...even if they know it's absolutely the right thing to do..they're like institutionalised prisoners...and yet there aren't enough of them, so very few of them need to worry about being fired or not being able to find another job...it's ironic because the engineers that have the balls to walk away and be confident in their actions are the ones that make the money and are desirable.
This is where a bit of enlightened cooperation can pay big dividends. If you are a large VMware user then club together with other large users and fund developers to add whatever you think is missing to open source VMware replacements. It might take a few months and some ££ but you then get what you want and save lots more ££ over years to come.
However it is psychologically hard for the large corporate C-suits to do this - they just see it as making payments when others, not in the paying club, get the results for free -- why should they do that ?
But that is how FLOSS works: some pay, some do not, all benefit.
It's a nice idea but we're stuck with a couple of problems.
C-suites like the idea of someone (i.e. a company) that they can shout out when things go wrong. They like the idea of guaranteed support levels and things like SLA's.
Time - time is really not on their side to do this. I can tell you from experience (see my comment above) that customers are looking and looking to do something fairly quickly.
Expense & complexity. It's likely that the features that I want, aren't necessarily the features that someone else wants or the ones they want and so on. Deciding what to prioritise comes at a cost in terms of financial and time and if someone feels they aren't getting the love they believe that they're paying for, then they're likely to walk away leaving others with higher bills. There's also the question of longevity - one of the selling points of virtualisation is the ability to run up virtual machines that may have been created on versions from many releases ago. Who's to say these FOSS offerings can provide that level of longevity?
Not saying it isn't a viable option if non of the above are show stopped, but for many, I think that they will be just that.
"They like the idea of guaranteed support levels and things like SLA's."
And exactly how have SLAs stopped Broadcom from shafting them?
They need to look on this as a learning experience. If they don't their competitors who do will be along to eat their lunch.
You can buy a support contract from Proxmox...what are you talking about?
Just because something is FOSS, it doesn't mean you can't purchase a support contract just like a commercial offering.
As for an SLA...that's why I exist. I provide the SLA as long as the customer provides me with the resources to carry it out...if a customer wants 99.9% uptime, they have to be willing to provide the kit to facilitate that kind of SLA...it's not rocket science.
They need to fire you and hire someone better. They aren't actually your customer if they're relying on vendor support contracts instead of you...you're just skimming in the middle.
The vendor support is for you mostly (which is why I'm usually happy if a customer is willing to pay for support), not really for them directly, it's a cheap way for firm to turn one techie into a team in an emergency and to ensure that there is knowledge on hand should you not possess it...I can't see the CEOs I work for calling a vendor at 2am for assistance...that's what they pay me for.
You're right but the market is being given a big incentive to fill the niche that Broadcom is providing. Assuming Proxmox or alternatives can achieve the right critical mass of customers they can start providing that kind of support – and it doesn't have to be much to improve on Broadcom's already pared down offerings: we had to wait over two weeks to get our initial VM Ware questions answered.
I can remember when open source, in any form, was considered too risky by companies but paid-for support for open source was probably really pioneered around Postgres (I don't think much of either RedHat or Canonical really) and has since spread to other areas. Support contracts with SLAs are now available and there are plenty of companies with track records over of ten years.
That said, I recently got an enquiry from NASA about my library and auditing, etc. Plenty of stuff is going to stay: use it at your own risk.
Advantage of FOSS is that, if you are not paying / paying much less licence fees, you can always pay for a couple of devs to make whatever changes / features / updates you need. And if you are a large org, you will probably come out ahead financially.
Forking hell! And before you know it you're catering to the whims of six different departments with customized forks and realize that your annual support and maintenance spend is more than it would have cost to buy VMware (the company, not the product).
Perhaps.
Or you pay whoever provides the support contract to do it. Or switch support over to someone who will.
Or you pay some outside consultant to do it.
The fact is that you have that choice, and so does everyone else using the open-source option so you get the benefit of the other businesses who did that
An easy way around this is just to adopt a new stack that is based on OSS tech even if it is "commercial". Proxmox is mentioned often, can pay them for their services, they appear to have a lot of open source stuff, and I assume they probably contribute upstream to projects. Same for Red Hat's Enterprise Virtualization, and I suspect HPE's latest hypervisor is similar. Maybe even Nutanix contributes upstream too for their AHV. (disclaimer I've never used any of these products)
And now that Veeam is oppering support for ProxMox natively for backups... even though people say ProxMox backup tools work great... it means one less thing to change in your infrastructure.
We've deployed some test instances of ProxMox on some ancient Cisco UCS hardware. Which works. It's different and I have to re-learn how to do somethings, and it's not quite as polished, but all I really want is NFS datastores, VMotion between nodes without having to tell my VMs they're moving (I'm looking at you Cisco for those crap ISE VMs which freak the hell out when you VMotion them!) and just work.
Don't really need much more than that. We're not that dynamic a shop, we don't need extran networking spun up, vritualized storage, or other things like that. Keep it simple, keep it reliable and let me get work done elsewhere.
This is the sort of response (the competition is Good Enough...) that should have the Broadcom suits in a collective coronary.
What *should* happen as small/medium operations take advantage of that is said competition expanding to cover the edge cases that are scaring off the bigger fish from following along: virtuous cycle?
It's not really re-learning per se...it's more like dropping proprietary terminology and substituting for the correct terminology and familiarising yourself with open standards and enjoying the fresh air and open space that comes with it.
With Proxmox you will never enter a meeting with a suggestion and wondering if it will require some kind of license or third party product. You'll be there to work out when to implement it, rather than how to budget for it.
Get ready for half the meetings and twice the productivity!
Quite right, Proxmox has pretty decent backup support as is, you can essentially do what you want the way you want it...you don't really need third party tools if you know how to write bash scripts and use rsync etc...if full VM backups are your thing.
Veeam is probably sweating because without VMWare their products aren't really that appealing...apart from migrating and converting VMs from VMWare to qcow2...which you can do without Veeam to be fair...I've done it tons of times and never once thought "You know what, it'd be way easier if I had a GUI with a wizard to do this".
It's the engineers out there that have spent a decade support VMWare products that I feel sorry for...they are about to cross over into a world where there are some seriously talented people that have been unrestricted for a long time and not limited by licensing that understand concepts and have implemented topologies most VMWare techies can only dream of.
I've worked with quite a few former "vendor locked" souls over the years...they're like deers in headlights when they have to change products...poor bastards.
Overspending on Cisco UCS kit is why you still have Cisco UCS kit.
I've never understood the whole Cisco thing...it's never been the best performing kit, nor the easiest kit to setup and they'd license the power button if they could...
I support where I need to, but I've never recommended it. You can usually get the same features, better performance and easier setup from elsewhere for a fraction of the cost.
I did a multi vendor test for a client of mine some time back when they were considering upgrading their network backbone to 10G...and across all the devices I tested (roughly 20 pairs of switches), Cisco never came out on top for anything other than sheer expense.
To my surprise, it was TP-Link that came out on top for most tests...throughput, general performance, cost, compatibility (with SFP modules and existing fibre plus PoE) and ease of use.
TP-Link was the first vendor I tried and because it was so damned easy to setup and hassle free, it made every other vendor look shit...it took me roughly 2 hours to complete all the tests on TP-Link kit...it took two days or more with Cisco (mostly because I had to keep requesting licenses) but also because certain things were fiddlier to setup on Cisco than TP-Link. Palo Alto was similar...but matched TP-Link in performance.
HP came pretty close to TP-Link...the only downer with HP kit is how noisy it is and how much power it draws.
Dell make some surprisingly good switches as well...not all of them are great, but certain switches in certain niches are pretty damned good...I have an 8 port Dell switch that I use for troubleshooting quite a lot because despite being a small and relatively cheap switch, it supports *everything* from VLANs to port mirroring etc...it's been a workhorse for me...I recently retired it in favour of an 8 port TP-Link switch (which also comes with PoE and is 2.5G) that does the same stuff and is equally as awesome, and I got it on sale for peanuts...but I'll always look at my Dell switch fondly because of the shit it got me out of...I'll probably still dust it off from time to time.
Where routers are concerned, it's never been a contest...in my earlier years, I was mostly deploying Cisco gear (back in the fucking Cisco PIX days, those were terrible and just generated arbitrary support tickets because of how fiddly it was to just create a port forwarding rule, you'd almost always have to go onsite because using a console cable directly was less dodgy than other methods), because that was the norm back then and I had to configure what was given to me...but for the last 10+ years it's mostly been pfSense / OPNSense...usually installed to replace a Juniper, Cisco or similar.
Draytek is still the king at SME level...their kit is everywhere and for what it costs I can see why...I don't particularly rate it that highly, and their firmware support leaves a little to be desired...but as far as customers are concerned, it just works and offers a hell of a lot for the money.
The reason this does not happen more often is 'money'....
In the long run , it will save money, but many C-suites/beancounters look at the fact that in the short term they don't believe migration work is productive, and therefore an unnecessary expense.
They are wholly incapable of looking at long term.
Meanwhile the techies who have to maintain the current POS are costing significantly more that the migration would in real terms. The C-suite can't see that though.
That's capitalsim for ya! The biggest problem is that politicians are basically corrupt and prevent competition authorities from doing anything until it's basically too late.
Competition authorities should make sure healthy competition remains in the market. Almost every market you can think of only has two or three incumbents and that's simply not enough.
I agree with your sentiment, but they can't just spawn competition into the market. There is a reason we don't see new search engines etc pop up and the one we do are just skins on the top of Google or Bing.
It's because to be up to speed with the incumbents you need the same money they're spending to be in the incumbents.
What competition authorities are supposed to do is ensure competition in emerging markets...they can't really do much about established markets...because they can't just pull billion dollar mega corps out of their asses.
The irony that this was below the article about Ransomware on the daily mail is not lost on me.
Can ‘Mark Lance, ransomware negotiator with GuidePoint Security’ be called in to negotiate with VMWare ?
—
As they allowed it, I think it’s EU/DoJ/FTC anti-trust territory…..
Worried about open source security?? I'd be more concerned about vSphere security because it's an absolute twat to update and a lot of people just don't bother...at least with ProxMox you can run a simple apt-get update && apt-get upgrade to get security patches...VMWare products have no such feature.
While you have a point, you won't be making friends with any managers who have to take the decision, who will rightly think you're sniping. Take their concerns seriously, show how they can be addressed with by the various solutions, including the incumbent and let them take an informed decision
I am an IT Manager, and the update mechanism is one of the things that would potentially swing my decision. Downtime for maintenance is something that is high up the list in terms of decision changing power...switching to maintenance mode, inserting an ISO and running an update script while all the VMs are offline...or just running apt-get then rebooting at a convenient time without taking anything offline for an extended period of time...hmm...
There are technical managers you know.
"...A few words:
Install Proxmox.
For corporates:
Install Proxmox, pay for support..."
The biggest problem here is that they only provide support between 9am and 5pm Austrian time (not Australian, as my phone so helpfully autocorrected it to last night without my noticing!).
And that is fine if it's all you need but I have customers that need 24/7 vendor support for when production systems die.
You have to be careful that what you're suggesting is fit for the market it's going into.
Yeah but chances are you installed Proxmox for high availability...because that is generally what it's known for, so a dead production server out of hours is part of a failover cluster right? RIGHT?
Fuck any business that wants 24/7 support on a single point of failure...I've got no time for bullshit like that.
< "The biggest problem here is that they only provide support between 9am and 5pm Austrian time..."
This is indeed an issue, but I'd wager that if even a small percentage of Broadcom customers told Proxmox that if 24/7 support were available, they would become Proxmox customers, then Proxmox would gladly expand the support team to accommodate them. If they have any foresight, they are already working on this.
I was an early bird to the VMWare space back in the mid 00's...I was pushing it hard to the MSP I was working at the time and I even got the VI3 certification eventually...but it never really paid off for me...that and they EOL'd my cert for no reason...I stopped supporting VMWare products when they stopped honouring my certification...they didn't even provide a path to upgrade my cert to the latest one for a cheaper price and a bit of a time save...the only option they offered at the time was to go through the whole thing again...training and everything. Pardon my French, but fuck that...they never even offered discounts for certified engineers to access their products...
...they also made dumbass decisions to prevent third party solutions to certain problems being practical...like backup software...it was always hellishly slow because SCP/SFTP was crippled and VMFS was shit.
VMWare in it's early days was amazing, but as soon as they started gathering pace and they developed a loyal following, they got greedy and shitty...as of about 2 years ago, I'd migrated all my clients away and moved to ProxMox...which to be fair is a better product in every way...the clients pay for the enterprise license for each of their setups (because it is massively cheaper than VMWare and is therefore a no brainer).
It's just a much easier product to support (especially if you're already familiar with Linux) because nothing is locked away or intentionally nerfed to make it unusable.
The only reason to deploy VMWare products now is if you're somehow locked in to their platform for some reason...because on a brand new project, it doesn't make sense at all...it doesn't have the features that the competition has (unless you pay for it), it doesn't have the performance and it doesn't scale without massive costs...it also doesn't run very well on exotic setups or setups that require hardware that isn't in the official "supported" list.
In 2024 VMWare products just feel clunky and old fashioned and you can't get the density that you can achieve with other products.
I have migrated my home stuff from VMware 6.5 to Proxmox on my Dell T320. VMware will no longer support my Dell T320 but Debian will and so will Proxmox because their distro is based on Debian. I now have OS supported gear again. That means I'm not throwing out my hardware - oooh that would be a green thing.
Now you will have to get a spreadsheet out to do the green calculus. An older box may not be so power efficient but I suggest that a server sent to the dump is probably worse than keeping it running.
I own a UK based IT company. The vast majority of my customers are running VMware, two run Hyper-V (bless).
We have already migrated one reasonably large customer from a three node VMware cluster with iSCSI SANs to a three node Proxmox n Ceph hyperconverged thingie. I've also "done myself" and a few others. The plan is to migrate all of our customers from VMware to Proxmox. We have a two year plan.
Even if Proxmox get the whiff of tonnes of cash from private equity and go mad, the VMs are QEMU/KVM - you just use another distro instead. Veeam will saunter along as required - they are jolly good at it.
Lovely 8)
The Hyper-V clients were probably former customers of a Microsoft Partner or something. I'll be a lot about their infrastructure is shitty. Having worked for a Microsoft Partner many moons ago, I can confirm that pushing Microsoft stuff comes ahead of picking the best technology, because there is massive profit in rolling out Microsoft licenses.
I also run my own IT business and I get a lot of MSP / Microsoft Partner refugees...originally it was people that had been saddled with SBS or similar products...these days the refugees want to get off VMware. Usually because they've had a scare and had to recover from backups, which takes forever if it's VMWare...they quickly learn that full virtual machine backups are less than ideal when it takes a week to get everything back up and running.
I've gone as far as bailing a business out by rebuilding everything from scratch and only recovering data because it is so much faster than restoring a VMWare VM.
If you look at the comments above, there's any number of users who've successfully told VMware/Broadcom to get stuffed and have moved to exactly the open-source-based systems you're pooh-poohing.
I'm sure there's some edge-case features in VMware that Proxmox/Nutanix/Red Hat aren't providing (yet...) but the extra business they're getting should let them build up to cover them. As it is it looks like the alternatives cover some large swath of those customers looking to jump ship.
Yeah that was back in the early days, and VMWare was even shittier. I remember being lured in by the possibility of simple BASH scripts to do backups...only to find that VMWare crippled the SSH client and you couldn't replace it...on purpose...because they wanted to shift other products.
VMWare spent a lot of time taking out third parties in the space as well. There was a time when Platespin was the product to have to move your VMs about...but it got killed off for a time, and I don't think anyone went back.
The thing that cracked me up the most was if you ran more than one VMWare box, you had to have a management box for it all, which had to be baremetal and they advised strongly against virtualising it...weird that.
For some of those options you mentioned to see nose to nose with VMWare, VMWare would have to up their game a lot...VMware is staring the competition in the balls these days and have done for a loooong time.
Simple shit like USB pass through is absolutely knackered on VMware...on Proxmox for example, I can use a single licensing dongle across essentially unlimited containers and VMs...on VMWare I can pass that same dongle (sometimes, if it works) through to one VM and one VM only.
I'm not saying that Proxmox makes "stretching" a license easier...but I've checked the license agreements on a lot of the products I have on client servers and there is nothing in the fine print about using one dongle across loads of containers...just that I can't use one dongle across many physical machines.
"28.6 percent are worried about open source security."
I burst out laughing when I read this. They're not concerned about security; they're worried because they won't have anyone to yell at when they're stuck.
If companies were led by IT people instead of professional managers, they would have chosen the most practical solution rather than putting all their trust in one option. This chaos exists because the C-level has been fixated on paying for SLAs instead of focusing on the product.the product.
The sad thing about is is mr dickhead Tan is going to win here....
They are greedy clever money men and they live for money, so will have done their sums well, guaranteed
They will have a detailed sheet for how much money they will actually make over the next five years or so from
1. cutting costs/innovation so that they only need to support the top x % of customers only
2. giving these top x% of customers all the left over support resources so that these customers feel safe albeit shafted
After 5 years ... who cares, sell it on to some dumb fecks to wring the last bit of monies out of it
Remember these pathetic sad arrogant rich vain people's might be total shitheads/parasite ..... but never underestimate their intelligence :)
I would substitute "cunning" for "intelligence" as in as cunning as a shithouse rat.
I would place "intelligence" between "cunning" and "wisdom."
Intelligent billionaires are about as rare as rocking-horse shit, wise ones non-existent or have some bloody small camels.
Shockingly, people with their entire IT infrastructure invested into a product that's basically abandoned them after a takeover and wants to charge them through the nose going forward... are considering changing.
I mean, the headline should be:
"Why the hell aren't the rest of the VMware customers are looking to move?"
Given Dockers recent announcement where they are lifting the cost of licencing by 50% ?? , people should be leaving in droves.but we know many won't.
It has become so ubiqutous that many devs can't see outside that box. So, alternatives like Podman still remain almost a 'side project' for many.
And we're a small shop with only 5 or so ESX servers. We just want a cluster of systems so we can VMotion stuff around at will and keep the lights on. We don't; need funky storage, VxRAIL, VxWhatever. Just a management pane, VMs, and hosts with NFS datastores we can shuffle things around on live. ProxMox looks pretty good for this, if only because the price is better.
But the C-suite always wants someone to blame or get support from, and they don't trust ProxMox quite yet. But with Veeam support, it's starting to look much better.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
