Things haven’t gone well for Broadcom since it acquired VMware, and much of what has happened has confirmed fears that Virtzilla customers expressed well before the deal closed.
Oh no! I guess.
Broadcom is out to juice a product for its value before throwing it in the bin. Broadcom knows exactly what it’s doing, and that is preying on customers and exploiting engineers.
I’m not even sure they’re going to bin it. Maybe shed the last-profitable customers to maximize profit from the large consumers. At least for a few years.
Call me cynical, but to me it’s about pushing consumers (SMB) to cloud providers instead of virtualizing on-prem with VMware. Broadcom can still sell VMware and support to larger vendors, or to SMB’s for a heftier profit margin.
I can only hope many SMB’s have support vendors who see the value in shifting to TrueNAS, Proxmox, etc (there are vendors who provide support contracts for those Open-source packages).
The problem with large customers is they can see value and if you charge too much it goes they can build their own.
I just left one of the UK’s largest VMWare customers, in a team working directly on VSphere and ESXI products. The team was just starting to tool up for a new internal project to manage VMs within the company using some of the newer features in these tools. We had well over 1 million VMs across more than a dozen datacenters. The cost of running this is expected to go up by 30 times or more.
While it’s going to take some time, they’re now looking at migrating to a different solution. So Broadcom are going to get their extra pay for a while, but not forever.
Not by next quarter. By the time that happens, this set of execs will be already at the next company.
“By next quarter” would be Proxmox
Often execs stick around longer than that.
Even if they do, they get bonuses for short term stock gain, not long term stuff. If the going gets bad, they totally do bail. Some companies - khm Reddit khm - even bring in execs just to take a fall when that happens, only for the same guys who set the problem up to take back the reins after the fallout.
They generally have a large part of their net worth in company stock and are getting options. Thus long term matters.
It’s supposed to work like that. You’re naive to think it actually does work like that for a majority of companies nowadays.
This sounds like it’s exactly what Broadcom intended. They are going to charge as much as they can and companies that depend on it will have to pay until they can move away and that may take years. Broadcom didn’t dig a hole. They triggered a trap on their customers.
Exactly.
Many SMB’s lack the in-house expertise to switch to something like Proxmox, and even then the support contract cost for things like Proxmox or TRUENAS, etc, isn’t much less than what VMware intends to charge.
Management in such businesses have a hard time comprehending the value in paying for a transition when the annual support costs at the moment don’t seem significantly lower (depending on number of processors, of course). There’s still the migration cost, and the risk of migrating.
SMB tends to outsource their system management a great deal - I can only hope these vendors see the value in migrating away from VMware.
What? Have you looked at the prices? Proxmox and TrueNAS have ridiculously low prices compared to VMware support. If we’re talking about Nutanix and stuff, sure - they aren’t cheaper than VMware. But Proxmox and TrueNAS are dirt cheap in comparison - not only because the documentation is pretty damn good, your standard, run of the mill Linux admin can do both, while Nutanix and the likes is an entirely different animal.
Full support from Proxmox isn’t cheap, compared to even the new prices on VMware, if you look at the per processor cost that small businesses often have.
As I said - it depends on processor count. I know a number of small businesses that will be paying $5k/year for VMware, not much more than Proxmox top tier (which is what they would want).
Proxmox is about $1500 per processor, so would be $3k-$6k/year for these businesses. That’s a trivial difference when you look at VMware already being installed and running, no transition costs, no risk of migration. You’d burn up a few $k difference with a single issue.
Then there’s backup - I doubt VEEM supports VM’s in Proxmox yet, and how do services like iLand work with Proxmox? These are associated costs, with migration costs and especially, risks.
Frankly, as much as VMware annoys the shit out of me, I couldn’t recommend migrating to Proxmox for those businesses, today. At best I’d recommend planning a transition when they need to upgrade servers, and do it early as a parallel install to give transition time for the business.
SMB doesn’t have the luxury of test labs for this stuff - they don’t have the cash flow/finance room to justify it.
Full support from Proxmox isn’t cheap, compared to even the new prices on VMware, if you look at the per processor cost that small businesses often have.
You’re joking, right? VSphere is AT LEAST 1400 per year for the base license, that hasn’t even got any support tickets - one Ticket is at least 300, 5 tickets is around 1200. Proxmox Full support starts at 340 Euros - with 3 Support tickets included. Then there’s also the fact, that Proxmox doesn’t have core limitations - meaning, you need at least two VSphere licenses for a 64-Core EPYC CPU. Oh, you want advanced networking or storage services? That’s even more.
As I said - it depends on processor count. I know a number of small businesses that will be paying $5k/year for VMware, not much more than Proxmox top tier (which is what they would want). Proxmox is about $1500 per processor, so would be $3k-$6k/year for these businesses. That’s a trivial difference when you look at VMware already being installed and running, no transition costs, no risk of migration. You’d burn up a few $k difference with a single issue.
WTF? You can’t even compare the 5k/year for VMware, just beacuse of the the single fact, that proxmox has UNLIMITED support tickets in the top tier. Not only that - it’s 1,1k per processor without any core limit - VSphere still has that ridiculous 32-Core Limit. In many cases, VMware also has support times up to 24 hours - proxmox has max. 2 hours
Frankly, as much as VMware annoys the shit out of me, I couldn’t recommend migrating to Proxmox for those businesses, today. At best I’d recommend planning a transition when they need to upgrade servers, and do it early as a parallel install to give transition time for the business.
SMB doesn’t have the luxury of test labs for this stuff - they don’t have the cash flow/finance room to justify it.
If they don’t, they don’t have the cash or finance room to justify their IT, period. For most SMBs, IT has become the utter lifeline for everything they do, that’s basically like when you are a machine shop without power. Meaning, the company is dead in the water for a serious period of time.
I’ve spent the last 15 years working with VMware exclusively. A little nervous about all of this
I was too, but as soon as I heard about the acquisition I started diversifying my non production kit for testing. I’ve now got Proxmox installed on an HPE DL380G10 with GPU pass thru, same on an HP Z440, and XenServer 8 installed on a pair of DL380G9 with MSA2040 backing storage.
At home I’ve got both truenas scale and truenas core set up each on a z230.
No matter what happens with the IT department at my office, I’m ready to either meet the new standards here, or go find work elsewhere.
When Broadcom announced the purchase a year or so ago, I abandoned all further VMWare certs, and put the time into getting my head around the alternatives.
I still have to use VMWare for 90% of my job, but I’m absolutely treating it like a locked-in platform, and assuming that anything I learn to do in VMWare, I need to understand the underlying concepts, not just their interpretation, and how I can do similar things on other platforms.I took a good look at Apache CloudStack. I of course have been migrating stuff to k8s and working with cloud native stuff that can be hypervisor independent or bare metal even.
I’m this close to an NSX-T cert and that will be the last one. That will be in high demand for quite some time I’d expect.
Is there not an alternative to what you use?
That’s not the issue in IT anymore. There are loads of alternatives, but rebuilding existing infrastructure on these kind of scales is nearly impossible without causing some serious downtime, loss of money or maybe even loss of life in case of some medical facilities.
You are in trouble my friend.
I’ve done this for so long I might just be like that old Cobol programmer in need during Y2K at some point.