Broadcom ditches VMware Cloud Service Providers Broadcom is tossing the majority of VMware's Cloud Services Providers as part of its shakeup of the virtualization titan's partner programs, say sources, leaving customers unclear who their IT supplier will be. The $61 billion purchase of VMware by Broadcom in November was swiftly followed by news of how it planned to …
I'm guessing these cloud providers use more sophisticated vmware software (probably vCloud and other stuff that I've never used) and Proxmox wouldn't be a suitable alternative out of the box (any more than ESXi+ vCenter by itself would be). They'd probably have to do quite a bit of custom work(provisioning/billing/resource management) which was probably why they decided on VMware to begin with to reduce the amount of work their staff had to do. Of course the biggest cloud players do all of that work anyway so don't need vmware.
They could probably try the OpenStack route but that would be a whole other can of worms with tons of new skills/staffing required in most cases, the small providers won't have the expertise to do it, and it may not be cost effective because of that.
sucks for them for sure though.
I doubt that with only weeks 15 remaining until VM pull the plug there's any reasonable chance of creating an operational alternative and migrating customers over.
In all likelihood this has fatally poisoned the Broadcom name even with the operators who receive the magic invitation letter, I’d certainly treat that letter as nothing more than an extended warning notice to change horses.
The late*, lamented Memset used to offer an OpenStack cloud amongst their products, and they were tiny compared to the big players. They also used Xen rather than VMWare for their more traditional VPSs.
*The name's still out there, but they're now just yet another Iomart brand.
Serious question: does Proxmox just replace the hypervisor, or does it have a good management interface as well? VMware brought virtualization to the masses by centralizing management and making it easy to use, whereas most open source projects spit contemptuously on the mere notion that usability might even be a concept worth momentarily contemplating pursuing (witness Kubernetes).
Answering my own question somewhat, it does appear that Proxmox is making an effort to provide a decent Web interface, so massive props to them for that. It appears that there may be other complications (e.g., no snapshot support on NFS, requiring Ceph or ZFS over iSCSI to be used for a clustered filesystem), but it's certainly an intriguing project. Just based on a cursory reading of the project page, it looks like you have to install Linux and KVM first, then you can deploy Proxmox, so it's a more heavyweight deployment than VMware Hypervisor (aka ESXi), but given that the commercial competitors have largely failed to steal any of VMware's thunder, perhaps Proxmox will succeed.
Hi,
you don't need to install linux and kvm first. Proxmox is a full distro deploying the full stack in 5min.
you can do snasphot on nfs of course.
The only limitation vs vmware is that you dont have vmfs, so for shared san storage (iscsi/fiberchannel), you use block directly (through lvm partition) and you don't have snasphot feature.
I think gfs2 cluster filesystem should come soon (some users already use it, but nothing official yeat), so it should be possible to do snasphot soon for san.
for all others storage (ceph, nfs, zfs over iscsi,), snapshots are available. (for block, Snasphot are done on the storage through their api . For file like nfs, it's done at file level like a vmdk)
Other missing feature is DRS, affinity/antifinity (on the roadmap for this year, some parts are already implemented).
network: you have advanced sdn features like nsx (with bgp-evpn, anycast routers, dhcp integration, firewall , ipam , ...)
ceph for HCI, zfs for local disk replication.
Proxmox VE has an okay management interface.
There are things you have to do on the CLI.
Probably the biggest thing preventing service providers from looking at it seriously is so many products that work with VMware do not work with Proxmox VE. (especially backup software).
I wouldn't count a tar file of the disk image files to be a proper backup solution (ie. built in Proxmox VE backup).
The networking is pretty simplistic, although I haven't really experimented with the new virtual network features of 8.1 much.
The API is pretty light weight.
One thing I encounter that bugs me to no end is that an NFS passthrough into a container doesn't "register" with the system for the first 5-8 minutes of uptime. Thus the container they are passed through to won't start in any reasonable time after host reboot.. I set a cronjob for 10 minutes post-boot to CLI bring up that container, and that seems to do the trick. Could just be my setup, but this seems to be consistent for me in multiple installs.
Kubernetes doesn't solve the same problem - running 100s of Windows VMs for corporate applications. Likewise Proxmox/Xen-Orchestra etc OK at some level of scale -neither can scale to a multi-customer platform?
Assume some/most the partners will get invites to new Broadcom scheme with new T's and C's attached - e.g. use Broadcom products over anyone else's'? They are Hostage to fortune now.
Also you can't just ask all your customers to covert VMs to containers overnight as a cloud provider. And even if you could - secure multi tenancy is bloody hard, so you still need that virtualization as you mentioned to split clusters up between customers.
Although - it's got me thinking - can you run a cluster with firecracker VMs as nodes: forming clusters of clusters? I'm about to do some looking into that. (Imagine DNS of the children would be a nightmare if it's possible?!)
Yes, but you can also do that with e.g. VirtualBox (which is still free to use even in commercial environments, as long as you avoid Oracle's VirtualBox Extensions poison pill). I've found VirtualBox works fine for my "just need a quick VM to do this" needs; I've run various Windows versions under it (including three Windows VMs running simultaneously on my laptop so I could have my own Windows domain and do some load-balancing testing), some SUSE versions, Kali Linux...
I've been saying the end of VMware has been coming for a long time.
An early straw for me was the Aug 12, 2008 time bomb they left in code. Sure...an accident, but certainly indicates they are more concerned about people NOT running their code than keeping it reliable.
Then there was the "let's replace our crappy bloated fat client with a web client" -- and managed to put out one of the last Flash-dependent applications pretending to be serious.
I recently had reason to set up a couple stand-alone ESXi systems, and I will admit, I was starting to think "maybe they are getting serious" -- the web host manager doesn't suck (other than the obvious, "We want you to buy vSphere" omissions -- I get that). Original plan was to do KVM, but that bogged down in "we got a VM running, but no one is really sure how", and moved to ESXi and been happy. But then this... "Dear customer...Make that former customer. Have a nice day".
People have wrapped good GUIs and web interfaces around KVM at various VPSs, so it looks like a good product. Just need some general guidelines to setting up a "big" environment, rather than the gobs of "hey, look, I set up a couple VMs on my dual-boot laptop, I'm an expert! Follow my advice, and here are 100 ads I'm getting paid for" guides that clog out the good info. A few good configuration tools (doesn't have to be web, can be console) would be nice.
And Promox is an alternative to VMware vs vSphere? A KVM in disguise?
And here we are talking about Cloud. To Promox reach Cloud environments as a Cloud product not as a Hypervisor, it needs to grow a lot. Promox is not an alternative to nothing.
Sadly, there is no good and reliable alternative to most of VMware products and environments, unfortunately. Because Broadcom should take a big lesson and lose a lot of money but unfortunately there is no alternative at the moment and they know this.
But Broadcom is doing the same that they did to Symantec and CA. There is no difference, the only difference is that VMware as a bigger market quota and makes more money than those 2.
...hear is my advice.
"Unfortunately VMware / Broadcom has ceased our contract with ourselves and many other providers with little or no warning, therefore your service has now ceased. You may wish to move to their service, where they may decide to cancel your contract in a similar manner and / or provide the service at a much greater cost. Our advice is to look elsewhere for your VM requirements, as they clearly don't care about you and will no doubt fuck you over again at some point in the future.
Yours, Cloud Provider"
I would wonder when in non US jurisdictions whether the communication and timing of the concellation ofthese contracts could be considered unreasonable and subject to equitable remedy.
This is the generic risk inherent in any externally provided service. Imagine AWS deciding to ditch all its bozo (below zero = less profitable) clients with the same timing and paucity of communication.
On-prem and bare metal off-prem as a service might be a tip for 2024.
Caveat Emptor.
Trusting 3rd party providers to underpin your companies existence should always be accompanied with an up to date get out of jail plan, this needs to cover 3 eventualities - 1: 3rd Party burns down with no backup* 2: 3rd Party give notice (such as here) for pulling plug. & 3: 3rd party hikes the cost in some manner because the think most users will suck it up.
* Being trashed by malware is the main threat.
very much this...
It stuns me when I see companies with decades or even a century or more history hitching themselves to 100% reliance on products put out by companies with just months or years of history without any kind of extraction plan.
EVERY SINGLE PART OF YOUR COMPANY'S OPERATION has to be replaceable. And in the case of IT, you probably want your old data on the new replacement system, so you better have a way (and a plan) to export your data to some other platform.
"Business continuity" is a real thing. But there appear to be many Boards of Directors who react in simplistic monkey-brain fashion. "No talk about bad thing! Talk about bad thing make bad thing happen!" (And the poo-flinging commences at the thoughtful person who raised the issue.)
Companies who had VMs on OVH's burned-up infrastructure have some relevant experience. It'd be nice to hear how those companies who did survive that disaster, did so.
"I would wonder when in non US jurisdictions whether the communication and timing of the concellation ofthese contracts could be considered unreasonable and subject to equitable remedy...."
Courts don't generally interfere with contractual agreements between companies, as it's understood that both parties knew what they were getting themselves into. "I didn't understand what I was signing" or "I didn't think I'd ever need to worry about that clause" aren't valid reasons to have parts of a contract set aside.
If your entire business model is built on your expectation of a continued relationship with a single vendor, you will be sorely disappointed.
It's a matter of WHEN, not IF, you get burned.
They won't. Because most of their small, less profitable customer require very little support. If they do require it, they either hire an MSP versed in AWS or pay AWS for it.
They already have the infrastructure, and it costs them nothing for these customers to run their workflows because the customer pays for it!
I've had an AWS account for years. I've never been charged more than $100 /yr because I only run things for very short periods while testing. And I get lots of credits from them for doing things like surveys.
I've never got any notice complaining I'm not making them enough money.
... and if they have plans for dealing with, "If any of our major material and/or service providers get a wild hair up their asses, jack up their rates, go out of business, and/or demand we wear Gooba-Jooba alien costumes before they'll sell their stuff to us".
If they do not have such plans, then those Boards of Directors have failed.
If your business is essentially reselling product X and (a) the makers of product X won't supply you and (b) migrating to product Y would be very expensive for you and your customers and (c) couldn't happen before X became unavailable, then your options seem rather limited.
@abend0c4: I was writing (less-than-clearly) about the BoDs of the end-users, not of the service resellers.
But if you're a reseller who resells only the product of Company X, if you're smart, you'll always be on the lookout for another ship to jump to, and migration plans for your customers to get to that ship.
I have worked for two small companies who resold only products of Companies X, and Y, respectively. If Companies X or Y had gone out of business, or created some crazy reseller rule or price change, both the companies I had worked for would probably have been ruined. (The first company had source code access to some of the apps we resold, but I don't have accountancy training, and thus could not have maintained those apps.)
Don't forget Broadcom acquired the 'King of Acquisition' company, CA (formerly Computer Associates). They're mainly mainframe products, and those products, by and large, have no other option than application replacement.
Anonymous as I have friends (mainly former colleagues from an acquired company) working there.
Seen a few suggestions for proxmox not a bad idea but in larger environments I’m not sure that would work. XCP-NG with XenOrchestra on the other hand might be worth a look at.
Gives a solid hypervisor along with a central management console!
Either way, looks like some of these smaller solutions might be about to get some more engineering resource if Broadcom royally p*** off the VMware dev teams
Neither one is a great choice.
Proxmox is great for a home lab but in production it has just too many bugs and strange issues. Overall it's more on par with Hyper-V than ESXi.
Xen is dead, the current version came out in 2010, and since then there has been little real development. On top of that, XCP-ng has inherited all the bad things from XenServer. Neither are great platforms to settle in 2024.
Neither of them are real vSphere alternatives. That's left to Nutanix, or maybe OpenNebula/OpenStack on KVM.
Why do Tech companies keep buying each other? I know that the founders can make a killing from it but in the wider picture it seems to killl good products off. I used to be a Solaris admin and welcomed the move Sun made to make it an open source OS just before the Oracle buyout but Larry killed that off because he really wanted Java. An alternative to Linux (yes I know BSD is there) would have been good but no.
Now Broadcom looks set to kill VMWare for all but the biggest customers. Is late stage capitilism simply borked or are tech billionaires arseholes?
Once I get a couple days free, I'm going to start work on getting a proxmox cluster built in a sandbox.
One of the problems with the entire VMware debacle, though, is that Cisco will not support any of their virtual appliances if they are not running on a supported hypervisor, and Proxmox is not one of those; This means that we'll be stuck with a couple ESX boxes just to run those for a while. :(
When a software company switches to a scammy subscription model, they raise rates every year or two. Eventually they lose most customers, but it doesn't matter. They have one or two locked-in victims who can't leave and are willing to be a trillion euros a month for the privilege of not having to replace their legacy platform. Just look at the mainframe for guidance.
It sounds like these resellers are a viable business, which is not a drain on Broadcom, yet BC want to kill off that whole channel -
reducing the installed base, losing the potential growth, and the ecosystem of vendors.
And on top of that, they don't care enough to come up with a exit/migration strategy, just to provide stability or as a reward for being paying customers and partners.
Weird.
I agree. Even if you really don't want the hassle of dealing with smaller customers, surely you partner with a large reseller and support shop or two you deal with in large concentrations and funnel them off to there?
Customer prices go up and you're splitting some of the profit, but it's better than this, surely?
Even at no profit broadcom's end (which obviously wouldn't be the case), it would keep people using VMware, for the skills market to not dry up, and prevent investment pouring into your competition when everyone jumps ship and keep you viable in the big customers you're trying to keep.
You're thinking too medium term.
This is entirely about the next six months at most. Next year is irrelevant, the manager will have moved on by then.
The incentives for upper management of publicly traded companies are almost never aligned with the long term health of the company they work for or employees they manage. Often directly opposed in fact.
Pump something, overhype something, bank the bonuses and then get hired away with a golden handshake before it comes crashing down.
While bonus clawback clauses are starting to exist since 2008 or so, I've yet to hear of a single case where they actually got invoked.
The Post Office seems to be the first time anyone has mooted doing so.
You're correct - the service provider program was run by "aggregators" for exactly that reason.
Not sure why Broadcom are doing this, as they were already charging a premium for the entitlement to run their software as a service provider, versus buying licences + support, when you looked at it over a 3 yr timeframe.
They also made it mandatory to use their service provider program if you were running shared platform services.....
... in preparation for shooting it with both barrels.
You could throw all sorts of shade at VMware for licensing costs but at the end of the day you bought a very good product that just did what it said on the tin. With a perpetual licence if that's the way you wanted to go.
I, personally, have two ESXi perpetual licenses. One for the server in my flat and another for a server hosted in NL, with the VMs essentially mirrored.
If Broadcom want a single penny from me... bring it on [smiles while, metaphorically speaking, polishes sword] ;-)
Unbelievable how channel unfriendly Broadcom/Vmware has become. Vmware is the anti-channel company....!! Remind of when Larry Ellison said "every time a partner makes money it takes a dollar out of my pocket". In the end I believe Broadcom/Vmware channel tactics will ultimately boomerang on them. VMware is already facing serious competition from cloud & container products like Redhat. So far Broadcom has acquired a company that is dependent on partners as VMware is. But for some odd reason Tan Hock views partners as competitors. In the end I believe that Broadcom will end up spiing off products or divesting them.
The MSP guy said it himself.
"How can they just cancel a major program affecting hundreds, perhaps thousands of customers..."
Not 10's of thousands of customers. Providing a service to small providers is fraught with difficulties for little profit. Better to deal with the big customers who require less support and generate bigger profits.
(The Ferengi would be proud of these guys)
Do I think it's a douche move? Sure!
I also do wonder why companies would want to deal with small MSPs. Sure, some of them are great, but most are not. Then dealing with them getting eventual acquired by a bigger company who suck more!
All these MSPs should remember Rule of Aquisition # 48 "The bugger the smile, the sharper the knife" and plan accordingly.
The one thing you can count on with a conglomerate take-over is they WILL leave thousands of people from employees to customers out to hang and dry, and think nothing of it as long as the shareholders are happy with the changes.
Remember, the modern customer is the shareholder; everyone else is just a user.
Broadcom just pulls the plug on the service on 01 April and leaves everyone out to dry. Perhaps then IT decision makers will get it through their heads that outsourcing your data to others means your business exists at the whim of someone who really doesn't care if you're still in business the instant it's no longer profitable or convenient to keep you as a customer.
This post has been deleted by its author
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
