Will Fusion be continued and supported?

[dminter]

@Technogeezer I direct this question to you as you have a great sense about these things.  Do you think Fusion will continue to be supported (and free for non-commercial users) as VMware has been acquired?  I know there was an official press release saying it would be...  But I am looking for your gut take.  It seems like such a trivial part of their enterprise, it's hard to see why Fusion would merit corporate resources. 

[Technogeezer]

I honestly have no idea. On one hand as you note there's the postings by Michael Roy that indicates that "Broadware" (VMware by Broadcom) is saying they will continue.

On the other hand, there are a stream of announcements like last week's cancellation of the free ESXi offering and the outright screwing over of both channel partners and perpetual license customers. And the absolute silence on an long-overdue update to Fusion to fix bugs that have been outstanding in Fusion since the Tech Preview - although to be fair, VMware didn't seem to have interest in fixing things before the Broadcom takeover either. The situation is much worse for the Workstation folks - there are some severe bugs/problems over there makes our bugs miniscule in comparison.

My gut tells me "trust for the time being, but verify by their future actions". I would not be surprised of  Broadware pulled a u-turn tomorrow and either kill the desktop virtualization products, remove the free options, or go to a subscription model. They've done nothing so far to earn my trust by their actions, no matter what Michael Roy has said.  I believe he's doing all he can, but there are decisions being made that are out of his hands.

Here's another gut take: "Broadware" is Computer Associates by another name, and so far they are living up to the well deserved CA reputation of "where software goes to die". I had hopes that the VMware folks would be able to change the outcome, but I'm afraid that like fighting the Borg in Star Trek, they're finding "resistance is futile". 

Personally, I'm hedging my bets. I'm continuing to use Fusion, but I've got UTM on my M1 Mac mini and have used it for macOS. I'm starting to play with it for Linux. It will also be interesting to see how Apple keeps evolving their underlying Hypervisor Frameworks in macOS, which UTM will be able to take advantage of. WWDC this year should be interesting. 

I've also got a Linux PC that's using the Linux libvirt, etc.

To me, if I don't see anything move in the next six months from "Broadware", I seriously have to think about making the switch off Fusion. There are plenty of low cost options for me without having to pay the Parallels tax.

[ColoradoMarmot]

I’ll chime in too.  I’ve been using Fusion since 1.0, and vastly prefer it’s tech and approach to the competition.

But we’ve seen Fusion continue to fall behind over the past two years.  We still are missing basic features like shared folders on Windows ARM, and graphics performance is really poor with certain apps.  Plus we have basic bugs like the ‘background service’ notification that haven’t been addressed - and all that’s raising serious concerns.

I know several of the Fusion team, and they’re all really good people, but it seems like corporate has been keeping the investment to a minimum for some time.  I don’t have insight into the financial case for continuing Fusion development, but I have to believe that it’s not a huge revenue generator - especially since they adopted the free for personal use model.  Given the new owners, that was a concern from the day the acquisition was announced.

That said, it was a solid gateway for people to learn the teach and start to interact with ESXI, so as a portfolio offering made sense.  But it’s clear by the elimination of the free esxi version, that the new owners are focused strictly on monetizing products.

I have real concerns about the future for Fusion.  It wouldn’t surprise me to see the free version go away, but I pay for mine so that’s not a personal concern.  Being tied to ESXI for things like tools and underlying tech brings both advantages and disadvantages, and on balance now, I’d have to say it looks like more an impediment than a help.  ARM tools is a perfect example of that.

I just had to set my dad up with a new ARM VM for running turbotax, and chose to do it on the competition because it supported shared folders, didn’t have some of the bugs, and had much better graphics performance.  I don’t like changing things on him, so that’s a permanent shift.

I figure if we don’t see some substantial improvements in Fusion, most notably bug fixes, shared folders and improved graphics performance by WWDC with the next tech preview,  I’m going to have to consider moving my own workload over to the competition.  That’ll be a sad day.

[gringley]

I sort of see a problem here like Microsoft had with Office on the Mac. Over the years Microsoft let Mac Office go its own way and Mac office became a very different product from Windows Office.  At some point Microsoft realized that as Mac market share grew having a version of the flagship product that was unique to Mac was bad for the brand and then Mac Office got redone to be like Windows Office.  Also having used Fusion since 1.0, and ESXi and Workstation over the years I see the same problem.  There are virtual hardware versions that seem sort of unique to Fusion and using Fusion is using a different product from the rest of the VMware ecosystem.  For me if Fusion is going to survive then Fusion needs to be very interchangeable with the rest of the VMware ecosystem to facilitate managing VMs for large groups of developers.  Parallels has won the war for the individual user market, and what it would take to get Fusion there would have little or no value in the enterprise markets Broadcom wants to live in.  

[Technogeezer]

@gringley you are on to something. I've used Fusion since V1 as well.

I preface all of this as my opinion... I could be wrong. 

What you say about Microsoft and the Office products is true. But Microsoft woke up. Their Mac software division has done a very nice job of supporting the Mac. Yes, they're not "real Mac apps", but they are light-years ahead on not making the Mac a second-class citizen like they did in the 90's. The current Office apps are very usable on the Mac and documents and skills can be switched between the Windows and Mac versions. And they are supporting macOS functionality. Their Mac software unit also provides Mac versions of other pieces of their software (Remote Desktop, OneDrive, Microsoft Defender come to mind immediately). Proving that even a battleship can turn on a dime given the right vision and resources. Let's also not forget the software that Microsoft is making available for iOS/iPadOS. If Microsoft can figure out how to have a non-Windows portfolio when faced with the Windows behemoth, one would think that VMware could do the same. Maybe spinning development of all their non-ESXi desktop products into a separate unit (outside of the control of ESXi) with their own resources would be a better idea.

Fusion absolutely is a unicorn in the VMware product line. It's the only one (save the ESXi on ARM fling) that runs on ARM CPUs, and the only one that to deal with macOS. It's likely to remain the only ARM product unless someone can convince VMware higher-ups that having a version of Workstation that runs on Windows 11 ARM is a good idea. Now there's an untapped market should non-Microsoft ARM PCs take off after the Qualcomm exclusivity agreement expires. (One would think they'd have a good idea on how to deal with virtualization on Windows 11 ARM because of the work they've done with Hyper-V on Windows and the "refactorization" of things they've had to do with native macOS and Microsoft hypervisors).

As much as VMware "borrows" from ESXi for the desktop virtualization products, Fusion (and Workstation) is being held back by that stance (you're right on point @ColoradoMarmot ). Compatibility with ESXi is not a selling point for Fusion unless VMware makes a real product out of the ESXi on ARM fling. However, I do understand that some of the "maturity" and features of the products are a result of ESXI "DNA". 

I would also think that VMware is not eyeing additional substantial investment in the Intel Mac platform. The inability to run old Intel macOS versions on Apple Silicon is already upon us thanks to Apple moving off Intel. And the end of the road is in sight for macOS support on Intel.  My tea leaves say that if you want to cut costs, Fusion support for Intel Macs is next on the chopping block. Would not surprise me if VMware freezes support for Fusion on Intel Macs in the not-so-distant future. 

[dminter]

You all  have excellent perspectives and make great points.  Anticipating the possible end of Fusion (and I need both Mac Intel as well as M1 support) - how hard are the non-paid alternatives to install and use Windows 11?  I don't need a step by step, just an idea.  My skills top out at using Terminal for basic stuff.  Thanks for all the POV's!

[Technogeezer · Accepted solution]

---

@dminter wrote:

Anticipating the possible end of Fusion (and I need both Mac Intel as well as M1 support) - how hard are the non-paid alternatives to install and use Windows 11?  I don't need a step by step, just an idea.  My skills top out at using Terminal for basic stuff.  Thanks for all the POV's!

---

My immediate reaction is to start weaning yourself off Mac Intel support sooner rather than later. Yes, there's software out there that won't run on newer macOS release (e.g. 32-bit apps). But those apps are abandonware if they have not been updated to run on newer macOS versions. And VMs to support that are only going to run on Intel Macs** which are going to get scarcer and scarcer. At some point in the future Intel Macs are going to stop receiving updates from Apple (making them a security risk).

I tried running Windows 11 ARM on UTM a while ago and it seemed to install OK. If I also remember from my earlier attempt, you needed to disable the TPM and secure boot checks in the Windows Setup, and also install the open-source SPICE tools (equivalent of VMware Tools), but they were available via a link from the UTM site and installed without much of a hassle. Note that you won't get all the bells and whistles of Fusion or Parallels, but for a lot of uses you'll probably be fine. 

I'll give it another whirl and report back as I think there have been changes in support of things like TPM and Secure Boot. 

** (that is unless the laws of physics are broken and the Intel chip emulation needed to run a virtual machine on ARM becomes as fast as native code. And no, "Rosetta can do it" because it doesn't. Rosetta recompiles/translates Intel application code (actually a subset of Intel CPU instructions) into ARM code. It does have a "dynamic translator" a.k.a. emulator for those instructions, but doesn't do that except in certain situations, and you'd know when that happens by the loss of speed. The tactics that Rosetta uses don't work very well for operating systems. And we all know how well QEMU Intel emulation works. So it's very unlikely IMO that we're going to see dramatic improvements in emulation speed that would make Intel VMs on ARM a viable option).

[ColoradoMarmot · Accepted solution]

First, as Technogeezer indicates, moving off intel mac sooner than later makes sense.  And there's no real option to run one VM on both.

But unless you're a hobbyist, Fusion or Parallels are really the best option for this.  It takes a lot of work to keep up with the changes that Apple keeps making, and having paid developers do that is far faster and more stable.

I don't think there's a rush to move off Fusion, unless (like my Dad) the lack of shared folders is a deal-breaker.  If it is, use Parallels.  If it isn't, then I'd hold tight and see what happens around WWDC timeframe.  It's not like Fusion is going to magically stop working in the next few months.

[dminter]

Much appreciated advice from all!  I am not a hobbiest but I hate the idea of paying parallels $100/year.  Seems like it used to be half that and I didn't look for a deal, so maybe it is still $50.

[dminter]

My Intel iMac is 5 years old so maybe I need to be thinking about a new computer anyway with the M series chip.  I am more a user than a hobbiest so I probably won't go down the DIY rabbit hole, but thank you for the perspective.  Much appreciated!

[dminter]

Thank you for the thoutful response @Technogeezer .  I think the UTM approach is more than I probably want to tackle. 

[Technogeezer · Accepted solution]

@dminter a followup on UTM running WIndows 11 ARM.

The latest version (4.4.5) will install Windows 11 ARM without a hitch. Looks like it has Secure Boot and TPM support, so there's no need to fiddle around with registry settings to get Windows to install. It also automatically mounts a VM to install the SPICE Guest Tools.

Two surprises:

Folder shaing (via the SPICE WebDAV driver that's installed as part of the tools) works out of the box.

If you run dxdiag in the VM, you find it's advertised as DirectX 12 - the Red Hat provided SPICE driver is supporting Direct3D DDI 12.0, and feature levels 12_0 and 11_1. All of these are nowhere to be found on VMware's drivers. 

One disappointment:

Sound support seems to be a bit wonky. It doesn't have the choppiness and lagging that we see sometimes in VMware, but it definitely isn't smooth either. "Static" is the best way I can describe it. Although things seem to have gotten better once I tweaked the sampling rate of the Windows sound output device to 44.1KHz instead of the default 48KHz.

But, it works and seems pretty snappy so far. Better than I would have expected.

[dminter]

@Technogeezer Very interesting and promising.  I don't need sound at all.  The only thing I run on Windows is Quicken.  Thank you for the further information!  I don't have a good feeling about Fusion staying around and knowing the options is very helpful

[Technogeezer]

Like I said, I'm hedging my bets and exploring options should VMware drop the hammer or continue to ignore Fusion. I'm not jumping ship yet, but VMware is running out of time.

[bryanjr]

Other than a short-term, painful, "guest OSes don't work with host OS VPNs", bug a few years ago, I've used Fusion for years and been very satisfied with it. I've always been a willing, paying customer because it just worked. If there aren't enough paying customers out there for it, it sounds like its days are numbered under Broadcom.

Does anyone have a guess on how many Fusion users there are out there in total? On Intel Mac or ARM (or both)?