What now for VMWare? What now for virtualisation?

For the most part this is largely not going to bother a company like Broadcom. They're at the stage nowadays akin to 'Nobody got fired for buying IBM' - expensive? Yes. A name change on products every 30 seconds? Also yes. But I've worked and consulted at some of the big players since BC took over, and none of them are going anywhere away from VMware. I'm talking 10k+ assets sorts of places. Lots of diversifying yes, but the reality is they don't need homelabbers and/or bloggers to keep the people they're targetting in the game. It's not popular, but it's the truth. What will more likely chase people away is the horrendous support if you don't have VCF licensing and you have to deal with the third parties.

They also have barely any rivals in the SDN space that hold a candle to NSX, agentless is such a big win for them. I will say though, VKS needs to die a death haha. That's another one that name changes every 30 seconds. I suspect they're pushing more heavily in the VCF/NSX/VKS/AVI space than just bog standard ESXi nowadays.

I won't be going back to ESXi in the homelab, not because I'm burnt by BC or anything, I work with NSX and the wider VCF BOM on a daily basis, but because I find it more interesting to work on open source tooling nowadays.

Not to say Broadcom are good or anything, they've dealt with a lot of things badly, they absolutely have and I don't think they're a good company at all. But people, especially bloggers/tech bro youtube folk seem to think their word is gospel, when the reality is the target audience for BC is super large customers.
We are reasonable sized enterprise and we have been working on exiting for almost 2 years and around 2 years left of our contract. Most of the peers I've spoken to are similar and even if the don't end up moving are seriously looking at alternatives and better still get of Windows / Linux and onto containers where they can. We are moving to a K8 / KVM solution and based on the number of people contacting me on linkedin its a fairly popular alternative choice.

BTW, we are NSX users and personally I could switch it off fast enough. Thankfully never had to use it in anger.
 
We are reasonable sized enterprise and we have been working on exiting for almost 2 years and around 2 years left of our contract. Most of the peers I've spoken to are similar and even if the don't end up moving are seriously looking at alternatives and better still get of Windows / Linux and onto containers where they can. We are moving to a K8 / KVM solution and based on the number of people contacting me on linkedin its a fairly popular alternative choice.

BTW, we are NSX users and personally I could switch it off fast enough. Thankfully never had to use it in anger.

I don't disagree, a lot of people are migrating - but reasonable Enterprise is again probably not their goal. I'm talking about places with 5k+ hosts, real sprawl and investment across the board. Using vRA, NSX etc. Most places aren't packing that much compute. The cost of them migrating is huge, from a vendor level, to process and time. Nobody at that level is paying anywhere near the price Broadcom are quoting either let's be honest. I'd be interested to see what the stats look like in three years, I suspect a lot of it's noise and token efforts and people will realise how good the actual Hypervisor product is compared to all the other stuff out there in terms of actual usage/administration effort. I say this as a Open Source fan too, I would deploy everything on KVM if I had my chocie, but alas the cost of Linux skilled engineers is higher than a vSphere admin.

Current place is migrating our K8s stuff on to VKS which I think is a huge mistake, we even engaged Broadcom PSO and that's been shocking, I've heavily recommended against it because it's a fundamentally bad product line with VKS/AVI. I'd have probably recommended Kubevirt. But it's significantly cheaper than AKS to do VMware.

I love NSX, I think it's great - but a lot of people hate it, I've found traditional Network Engineers tend to hate it. Our Network team wants nothing to do with it. Whereas I've done it for a number of years now.
 
I was thinking that I might not need to run a full OS in a VM, just for the following that I have..

Plex Media server, one VM for this alone.
FTP Client and DC++ on another VM.
'Torrents' VM with Nord on it.
DC++ Hub on a VM

I also run a Pi5 with PiHole on it, which could go elsewhere not on physical hardware.

I run pretty much all those as LXCs, even Qbittorrent with NordVPN.

In fact the only thing I run as VMs are Proxmox Backup Server, a Windows VM I use mainly as a remote client so I'm not booting up my gaming machine just to browse the internet, and finally Home Assistant - although there is a container version you can deploy, I've just found it more of a pain to use than a VM.
 
I don't disagree, a lot of people are migrating - but reasonable Enterprise is again probably not their goal. I'm talking about places with 5k+ hosts, real sprawl and investment across the board. Using vRA, NSX etc. Most places aren't packing that much compute. The cost of them migrating is huge, from a vendor level, to process and time. Nobody at that level is paying anywhere near the price Broadcom are quoting either let's be honest. I'd be interested to see what the stats look like in three years, I suspect a lot of it's noise and token efforts and people will realise how good the actual Hypervisor product is compared to all the other stuff out there in terms of actual usage/administration effort. I say this as a Open Source fan too, I would deploy everything on KVM if I had my chocie, but alas the cost of Linux skilled engineers is higher than a vSphere admin.

Current place is migrating our K8s stuff on to VKS which I think is a huge mistake, we even engaged Broadcom PSO and that's been shocking, I've heavily recommended against it because it's a fundamentally bad product line with VKS/AVI. I'd have probably recommended Kubevirt. But it's significantly cheaper than AKS to do VMware.

I love NSX, I think it's great - but a lot of people hate it, I've found traditional Network Engineers tend to hate it. Our Network team wants nothing to do with it. Whereas I've done it for a number of years now.
We aren't huge but 3,500 hosts running 5,000 applications with 25,000 VM across 200 clusters using NAS from 100+ host clusters in datacenters running NSX + SRM to two host branch offices with local storage. I agree about network engineers not liking NSX but we have had stability problems, some of it VMWare issues, some it Cisco. But to be honest implementing ACI SD-LAN has been a PIA though now its stable we don't need NSX any more and our target platform migration tool didn't work with NSX anyway so it had to go prior to migration.

Annoying thing is that VMWare is pretty mature and has loads of support, pretty much all our COTS apps will work on it but not may certify on KVM. I like Aria, again a good product and like most companies we've invested thousands of hours of development into automations. And there is a wide pool of engineers available. KVM is baked into Linux so there is a reasonable of usage and knowledge. Proxmox, Nutanix, OSV, Citrix and pretty much all the alternate hypervisors use it. The only real alternative to ESX and KVM is Hyper-V which actually I am a fan off for SMB environments.

Overall, its all about the management tools, automation, etc. I've noticed trend to converge Container hosting and VM into the same management plane, something that VMware do.

We had a look at Tanzu when it came out as we were just getting into on-prem K8 at the time but no chance I would recommend it. Went with OpenShift and while it has its flaws, after around 4 years of production usage its stable and we've built up a good amount of experience in it now.
 
I run pretty much all those as LXCs, even Qbittorrent with NordVPN.

In fact the only thing I run as VMs are Proxmox Backup Server, a Windows VM I use mainly as a remote client so I'm not booting up my gaming machine just to browse the internet, and finally Home Assistant - although there is a container version you can deploy, I've just found it more of a pain to use than a VM.
I see Plex and DC++ both have docker versions and there are torrent clients available as docker images.

Oddly FTP is a bit tricky. As it relies on being able to open range ports, that isn't something contains and docker are good at.
 
We aren't huge but 3,500 hosts running 5,000 applications with 25,000 VM across 200 clusters using NAS from 100+ host clusters in datacenters running NSX + SRM to two host branch offices with local storage. I agree about network engineers not liking NSX but we have had stability problems, some of it VMWare issues, some it Cisco. But to be honest implementing ACI SD-LAN has been a PIA though now its stable we don't need NSX any more and our target platform migration tool didn't work with NSX anyway so it had to go prior to migration.

Annoying thing is that VMWare is pretty mature and has loads of support, pretty much all our COTS apps will work on it but not may certify on KVM. I like Aria, again a good product and like most companies we've invested thousands of hours of development into automations. And there is a wide pool of engineers available. KVM is baked into Linux so there is a reasonable of usage and knowledge. Proxmox, Nutanix, OSV, Citrix and pretty much all the alternate hypervisors use it. The only real alternative to ESX and KVM is Hyper-V which actually I am a fan off for SMB environments.

Overall, its all about the management tools, automation, etc. I've noticed trend to converge Container hosting and VM into the same management plane, something that VMware do.

We had a look at Tanzu when it came out as we were just getting into on-prem K8 at the time but no chance I would recommend it. Went with OpenShift and while it has its flaws, after around 4 years of production usage its stable and we've built up a good amount of experience in it now.

ACI - thanks for the PTSD chief. I thought I'd scrubbed that horror story out of my mind. I inherited what TAC called the worst ever ACI deployment they'd ever seen which was fun. We have issues with our current NSX which use the old NSX-T LB's and our Edges constantly run out of memory because we do load balancing very, very wrong which I am trying to fix at the moment.

Agree 100% with Tanzu too. I think the whole product line is a mess, they've tried to simplify it, but so much of their own documentation still refers to VKS as TKG and it still keeps people confused because there was at one point it felt like aobut 6-7 products under Tanzu.

I truly think that Nutanix would have made huge gains if they'd supported Block, as it's not a terrible product. Although Pure are attending a Nutanix event and saying there's a big announcement, so it would be interested to see what happens there - hopefully not just FlashBlade support.
 
ACI - thanks for the PTSD chief. I thought I'd scrubbed that horror story out of my mind. I inherited what TAC called the worst ever ACI deployment they'd ever seen which was fun. We have issues with our current NSX which use the old NSX-T LB's and our Edges constantly run out of memory because we do load balancing very, very wrong which I am trying to fix at the moment.

Agree 100% with Tanzu too. I think the whole product line is a mess, they've tried to simplify it, but so much of their own documentation still refers to VKS as TKG and it still keeps people confused because there was at one point it felt like aobut 6-7 products under Tanzu.

I truly think that Nutanix would have made huge gains if they'd supported Block, as it's not a terrible product. Although Pure are attending a Nutanix event and saying there's a big announcement, so it would be interested to see what happens there - hopefully not just FlashBlade support.
Nutanix as a product isn't bad, the issue of cost of storage. We did some calculation vs NAS and it was coming in at double the price.
 
Nutanix as a product isn't bad, the issue of cost of storage. We did some calculation vs NAS and it was coming in at double the price.

Yeah, at my last place we quoted replacing VMware with Nutanix, but the problem was we had between 50-60 block SANs and a fair bit of additional NAS, it simply wasn't even remotely economical.

I'll look at LXC's, any specific hubs for learning these?

Always start with YouTube

I prefer, always start with documentation. But I despise the general 'tutorialgate' of Youtube where they do intros, then some vague sponsor, then a beg for likes and subscribes, then eke out a video for as long as possible to hit engagement for money etc. But I'm grumpy with 'tech influencer yotubues giving horrible info and bad takes' these days.


I'd highly recommend Incus over LXD if you want that centralised management, Canonical screwed LXD up and the guy developing Incus quit and forked the good LXD pre-license change and improved it. Both use LXC.

---

Alternatively, if you really want to have fun, move over to BSD and use Bhyve or Jails or VMM/VMD.
 
Last edited:
always start with documentation
If only the big enterprise customers would do that, rather than ignore it, get into a bunch of trouble and then blame the vendor for not telling them. :confused:

Amen on YouTube. Not that there's a lack of decent content, but the vast majority is clickbait/sponsored dross. People who have very little knowledge/experience in areas of technology making some wild and unsubstantiated claims/statements etc and many sheep follow. Funnily enough, the folks with real experience - the CCIE/VCIX's of the world etc - don't have the time to record proper content, edit it, and run a channel.

Think I'm done here, can't deal with the FUD and scaremongering. :cry: The ACI references tipped me over the 'edge'. Badum tss!
 
If only the big enterprise customers would do that, rather than ignore it, get into a bunch of trouble and then blame the vendor for not telling them. :confused:

Amen on YouTube. Not that there's a lack of decent content, but the vast majority is clickbait/sponsored dross. People who have very little knowledge/experience in areas of technology making some wild and unsubstantiated claims/statements etc and many sheep follow. Funnily enough, the folks with real experience - the CCIE/VCIX's of the world etc - don't have the time to record proper content, edit it, and run a channel.

Think I'm done here, can't deal with the FUD and scaremongering. :cry: The ACI references tipped me over the 'edge'. Badum tss!
Agreed. Anyone that has the time to do a full time job and run a YouTube channel, well, doesn't really have a full time job. The same goes for chronic bloggers/LinkedIn posters/content creators/so on. Their blog/channel/etc is what's actually their full time job.

In other news, we are still struggling to find a suitable replacement for VMware in our estate due to the specialist appliances we run that underpin our platform. They have us by the balls. Still.
 
I had coffee yesterday with the ex-CIO of VMware - he left the moment the Broadcom deal was signed. Most key vmware staff saw the writing on the wall, knew what Broadcom ownership meant for the company and their customers and left at the first opportunity.

Such a shame; I’ve been a VMware fan for a long time, but couldn’t in good conscience give them my money now.
 
Got stuff in a client's estate running on VMware currently, they're planning to move to Nutanix apparently... apprehensive to see how that works out for us :/

Interestingly, Nutanix have just announced they'll be supporting Pure FlashArrays which is a big win for them really - if only they'd done it pre-BC buyout they'd have probably taken quite a lot of wins from that, feels a bit late.
 
With V9 being VVF & VCF Only - no longer a seperate Standard or Enterprise Plus edition, I can see a lot more of the customer base accelerating their move away from VMware/Broadcom. I work at a VAR and the conversation is daily with people looking into their options, but I now I fear there is not an option for most small to medium environments wanting to use vSphere.

Pretty sad for me, my career has been built around VMware based solutions for the last 13+ years.
 
With V9 being VVF & VCF Only - no longer a seperate Standard or Enterprise Plus edition, I can see a lot more of the customer base accelerating their move away from VMware/Broadcom. I work at a VAR and the conversation is daily with people looking into their options, but I now I fear there is not an option for most small to medium environments wanting to use vSphere.

Pretty sad for me, my career has been built around VMware based solutions for the last 13+ years.

Yeah let's be honest they're not interested in them unfortunately. They're interested in the big players who are all so tightly integrated with things like NSX and vRA. All the big boys, whilst diversifying as they always do aren't moving anywhere fast from VMware. I did some work with one of the big consultancies during the BC acquisition and overall probably one of the biggest accounts UK across the customer base and they pretty much immediately said not a chance they would move from VMware - I imagine the costs there though they're not paying anything close to what SMB VVF are being charged.

Just done another couple of VCF deployments myself, more frustrated at the fact that vvols got wiped out with basically no notice and had to hear it from the storage vendor leaks. Good decision to move away from vSAN requirements too in the management domain, just as we've bought a bunch of vSAN nodes, we'd have saved a tidy sum if 9 had been GA a few weeks earlier haha.

Used to work in channel, and wonder how many of the big boys are direct with them now - I specialised in the VCPP space so always had direct for the most part. I'm with you though, been a VMware and Storage specialist and pretty integrated with NSX nowadays, but took my last role assist with VMware but primarily do Linux/Automation just so I don't have all of my eggs in the one basket.

Also if they rename vRealize one more time, I'm gonna slap someone at Explore with a great big dirty fish.
 
Back
Top Bottom