Broadcom to buy VMware in $61 billion acquisition

[jlbenedict]

Goodbye VMware... it was good knowing you.

https://www.reuters.com/article/vmw...om-to-buy-vmware-in-61-bln-deal-idUSL3N2XI2A1

"Broadcom Inc said on Thursday it will acquire cloud computing company VMware Inc in a $61 billion cash-and stock deal, the chipmaker’s biggest and boldest bid to diversify its business into enterprise software.
The acquisition is the second biggest announced globally so far this year, trailing only Microsoft Corp’s $68.7 billion deal to buy video game maker Activision Blizzard Inc.
The offer of $142.50 in cash or 0.2520 of a Broadcom share for each VMware stock represents a premium of nearly 49% to the stock’s last close before talks of the deal were first reported on May 22. Broadcom will also assume $8 billion of VMware’s net debt."

 

[Zarathustra[H]]

I'm surprised VMWare is even as big as they are anymore.

They were the only game in town in the beginning, but already back maybe ~8 years ago it felt like everyone, even big enterprise users, were in the process of transitioning away from VMWare to alternatives like KVM.

 

[Lakados]

I'm surprised VMWare is even as big as they are anymore.

They were the only game in town in the beginning, but already back maybe ~8 years ago it felt like everyone, even big enterprise users, were in the process of transitioning away from VMWare to alternatives like KVM.

The thing is, once you are in and using VMWare it is not the easiest to transition off, they do some fun stuff with their vHD formats that make it... problematic to simply transition over without fully rebuilding the systems and depending on the software you are running it really messes with the licensing. Microsoft released a helpful set of tools that let you convert them over to a format that works with Azure of Hyper-V. However, you still need to take note of some specifics like MAC addresses and other hardware identifiers that may have been latched onto by software for their licensing systems.
VMWare still does good work but Dell hasn't done a lot with it in the last 7 years, I mean dell has done a lot with EMC's storage technology, but they sort of left VMWare to linger and it shows.
I still have one remaining VMWare system and we needed VMWare, specifically for how it deals with virtualized GPUs for a remote workstation pool.
But going forward I would not have used them if it weren't for Dell basically giving me the licenses with the server purchases, looking at alternative solutions from companies like Scale do a better job and are easier to manage as they have actually kept up on development and UI improvements.

 

[4saken]

With AWS/GCP and soon Azure, working with VMware Cloud Engine in the hyperscalers, makes a lot of sense for companies who want to lift and shift workloads to a cloud. Broadcom is basically a tech holding company at this point anyway, so I don't see much changing. Those legacy/monolithic apps can now just be vmotioned easily to a cloud, and since so many enterprises/orgs have these initiatives, should be interesting to see how much longer VMware can still say relevant. I think its a stop gap in the cloud until organizations retire those technical debt applications and move to first party services on AWS/GCP/etc.

 

[jlbenedict]

Cloud and containerization seems to be the future (at least for a bit until something else comes along).
Writing was on the wall as soon as it came to be that there wouldn't be a stand-alone release of Hyper-V 2022.

 

[Lakados]

With AWS/GCP and soon Azure, working with VMware Cloud Engine in the hyperscalers, makes a lot of sense for companies who want to lift and shift workloads to a cloud. Broadcom is basically a tech holding company at this point anyway, so I don't see much changing. Those legacy/monolithic apps can now just be vmotioned easily to a cloud, and since so many enterprises/orgs have these initiatives, should be interesting to see how much longer VMware can still say relevant. I think its a stop gap in the cloud until organizations retire those technical debt applications and move to first party services on AWS/GCP/etc.

Cloud services get expensive when GPU workloads are included, that's where VMWare's bread and butter is. At this point as it works very well with Nvidia's vGPU hardware, Nvidia and Dell/EMC worked pretty closely on it and it just works. So if you need to have one box or a stack that is farming out remote workstations VMWare Horizons is about the best approach for that right now, but there are a number of other companies nipping on their heels as Dell/EMC let the tech linger too long and lost their edge.

 

[Sycraft]

I'm surprised VMWare is even as big as they are anymore.

They were the only game in town in the beginning, but already back maybe ~8 years ago it felt like everyone, even big enterprise users, were in the process of transitioning away from VMWare to alternatives like KVM.

They still make the best, most complete, hypervisor solution. We use Hyper-V here, licensing costs got us to change mostly, but VMWare really is a cut above. I still use it for desktop stuff and play with the enterprise stuff occasionally and particularly if you want to do a mixed OS environment it just kills. It is expensive, but for some the cost isn't an issue, it is worth it.

Cloud and containerization seems to be the future (at least for a bit until something else comes along).
Writing was on the wall as soon as it came to be that there wouldn't be a stand-alone release of Hyper-V 2022.

Yes and no. While plenty of businesses went wild shoveling everything out to the cloud as fast as they could, others have not. Also some of those that did are discovering a few things about the cloud:

1) The cloud is only cheap if you don't use much. There's a reason why MS and Amazon have such stupidly large profit margins: They charge the fuck out of you for cloud service. If you need a system full time, it can easily run you the complete cost of a server in 2-4 months to have it running. So it can be cheap if you have a workload that scales up and down real fast and you can keep it lean and mean, but that isn't the case for a lot of things. While paying monthly masks the true costs against people who don't know how to annualize things, eventually there is some executive smart enough to do that and say "We are paying HOW MUCH for this?"

2) Latency/availability can be a real issue. Shovel everything out to another data center, and suddenly there is latency that wasn't there before. This can be an issue for some things. So you start to see some silly things like "edge cloud" which means "Put a server on prem, that you pay for, but pay a cloud provider to manage it."

3) Some people are wondering if maybe giving all your data to Amazon, a company not known for being super ethical, is perhaps not the best idea. Sure they claim they won't be looking at your shit for trade secrets, but you have only their word on that.


At any rate, some companies are not as interested in the cloud, or only want it as part of their solution. For them, VMWare may be an answer. They have some slick stuff if you want to do things like VDI where you build up an on prem solution that meets your normal load, but have cloud availability to deal with peak load. That way you aren't playing out the ass for stuff you are using all the time, but if you don't have to have a bunch of extra hardware sitting idle 99% of the time to meet the occasional demand spike.


Always remember that just because there's a new, shiny, trendy solution it doesn't mean the old solutions necessarily go away. One of the examples I love is tape storage. Not only is it still used, it is still under active development. We just bought a brand new LTO-9 tape drive. Sounds archaic, and there are things about tape that suck... but it is $5/TB for data storage and that's a one-time cost and scales basically infinitely. You can packrat data at a huge scale for a price way less than you are going to get with newer technologies.

 

[Zarathustra[H]]

The thing is, once you are in and using VMWare it is not the easiest to transition off, they do some fun stuff with their vHD formats that make it... problematic to simply transition over without fully rebuilding the systems and depending on the software you are running it really messes with the licensing. Microsoft released a helpful set of tools that let you convert them over to a format that works with Azure of Hyper-V. However, you still need to take note of some specifics like MAC addresses and other hardware identifiers that may have been latched onto by software for their licensing systems.
VMWare still does good work but Dell hasn't done a lot with it in the last 7 years, I mean dell has done a lot with EMC's storage technology, but they sort of left VMWare to linger and it shows.
I still have one remaining VMWare system and we needed VMWare, specifically for how it deals with virtualized GPUs for a remote workstation pool.
But going forward I would not have used them if it weren't for Dell basically giving me the licenses with the server purchases, looking at alternative solutions from companies like Scale do a better job and are easier to manage as they have actually kept up on development and UI improvements.

Hmm.

I don't recall having any of these issues when I transitioned my home ESXi server to a combination of KVM and LXC containers 6 years ago. but that said, I wasn't using any servers or other software that depended on licensing. It was mostly open source stuff.

I found it fairly trivial to convert the VMDK's to raw drive images using a command line tool (qemu-img I think? I can't remember, it's been a few years after all) and then importing the raw drive images into KVM or LXC.

But again, my 20-some guests were all BSD or Linux based and running largely open source software. The only software that was not open source was closed source and proprietary, but not licenced in that way, most provided with hardware (like Unifi controller software)

I can totally see how MAC addresses moving around causing licensing issues. Each hypervisor emulates different hardware too, so switching from one to another could result in triggering a change to system fingerprint style security, like that Microsoft uses for Windows.

My migration - while huge and complex for a home system - was likely much more simple than most enterprise implementations though.

It's just that at the time, it seemed people were migrating to KVM in droves, both home lab users and enterprise users. I guess I just figured VMWare was diminished in the marketplace after that.

 

[Lakados]

Hmm.

I don't recall having any of these issues when I transitioned my home ESXi server to a combination of KVM and LXC containers 6 years ago. but that said, I wasn't using any servers or other software that depended on licensing. It was mostly open source stuff.

I found it fairly trivial to convert the vdi's to raw drive images using a VMWare command line tool (can't remember the name of it) and then importing the raw drive images into KVM or LXC.

But again, my 20-some guests were all BSD or Linux based and running largely open source software. The only software that was not open source was closed source and proprietary, but not licenced in that way, most provided with hardware (like Unifi controller software)

I can totally see how MAC addresses moving around causing licensing issues. Each hypervisor emulates different hardware too, so switching from one to another could result in triggering a change to system fingerprint style security, like that Microsoft uses for Windows.

My migration - while huge and complex for a home system - was likely much more simple than most enterprise implementations though.

Yeah for me it really messed with my licensing for my Mitel systems, didn't like my Ricoh stuff either. Ours were all set up with the VM's stored on a storage server and the hosts connected back with iSCSI and that seems to do things differently than if you have it stored on the same box that processes it. It was not a fun transition job. I am glad to be mostly free of VMWare.

 

[Balkroth]

VMWare Telco Cloud is going to be a huge thing for ORAN/VRAN .

They're still the best for a lot of solutions, and ease of use.

 

[Sycraft]

It's just that at the time, it seemed people were migrating to KVM in droves, both home lab users and enterprise users. I guess I just figured VMWare was diminished in the marketplace after that.

You have to remember that is the stuff you hear about, you don't tend to hear about the stuff that doesn't change. When there's something new and trendy you'll hear lots of chatter about it, but chatter doesn't necessarily translate to what everyone is doing, just what the chatty people are doing.

An example I like for that is programming languages, specifically C/C++. Periodically there will be some report that talks about what languages are popular and being used and invariably those two don't make the list. Yet it takes only a little research to determine that's still what most software out there is written in, particularly when you start to consider all the embedded and specialized devices in the world. So what gives? Well what gives is they base this off of what people chat about online, or what people responded to a survey, not any sort of actual analysis. The people/companies that do C and have been doing it for decades aren't chattering up a storm like the people who've just discovered their new favourite language of the month, but that doesn't mean it is gone.

Similar deal here I think. People who were changing to KVM wanted to chat about it, people who were sticking with VMWare had nothing to say.

 

[D-EJ915]

KVM is a fine virtualization tech, I mean I used it 10 years ago with no issues but it's not something you can sell since it's not an actual product and thus doesn't have the same support or integration.

 

[ND40oz]

KVM is a fine virtualization tech, I mean I used it 10 years ago with no issues but it's not something you can sell since it's not an actual product and thus doesn't have the same support or integration.

Yeah, I think I'd go Nutanix or Hyper-V before I'd go KVM if I were to migrate to another hypervisor. The support just isn't there for it. You have to think full integration including backups and DR, things like desktop and app virtualization.

 

[bigdogchris]

VMWare and their support is awesome. Love the product.

 

[YeuEmMaiMai]

61 Billion,lol that's stupid money right there... wow

 

[cjcox]

Well, now at least when there's a network problem in VMware, we can all say, "oh yeah, Broadcom..."

 

[OFaceSIG]

Yeah, I think I'd go Nutanix or Hyper-V before I'd go KVM if I were to migrate to another hypervisor. The support just isn't there for it. You have to think full integration including backups and DR, things like desktop and app virtualization.

The Nutanix file system for the underlying storage is unreal.

 

[LodeRunner]

VMWare and their support is awesome. Love the product.

Yeah, at a previous job I managed a VMWare cluster and the support I got for SEV1 and SEV2 issues was great but response times and helpfulness started getting worse 2017 or so. I don't deal with VMWare anymore, but I keep hearing how their support is lackluster these days.

Still a rock-solid product; the stuff I had to open tickets for was some weird edge-case stuff in the VCSA; never had an actual problem with ESXi itself.

 

[cjcox]

My recent VMware support is like:

"Did that work?"

No.

"How about now?"

No.

"Now?"

No.

"Try this. Now?"

No.

"I don't see how this will help, how about now?"

Yes... seems like it's working.

"Hmmm... ok... I'm going to mark the ticket closed, make sure you give us a great rating."

 

[cjcox]

 

[Lakados]

My recent VMware support is like:

"Did that work?"

No.

"How about now?"

No.

"Now?"

No.

"Try this. Now?"

No.

"I don't see how this will help, how about now?"

Yes... seems like it's working.

"Hmmm... ok... I'm going to mark the ticket closed, make sure you give us a great rating."

That sounds about right, I'm waiting for them to answer because now when I try to log into the hypervisor it kicks me out saying my licenses are invalid...

 

[UnknownSouljer]

View attachment 477558

Can't be any worse than them essentially rotting on the vine at Dell.
But perhaps neglect is benevolent compared to active mismanagement.

 

[jlbenedict]

If VMware support through Broadcom is anything like Symantec support......

 

[ElementDave]

Always remember that just because there's a new, shiny, trendy solution it doesn't mean the old solutions necessarily go away. One of the examples I love is tape storage. Not only is it still used, it is still under active development. We just bought a brand new LTO-9 tape drive. Sounds archaic, and there are things about tape that suck... but it is $5/TB for data storage and that's a one-time cost and scales basically infinitely. You can packrat data at a huge scale for a price way less than you are going to get with newer technologies.

Interesting example. I'd all but forgotten about magnetic tapes. I knew they weren't extinct, but didn't realize they were still actively developed.

Why the Future of Data Storage is (Still) Magnetic Tape

In 2015, the Information Storage Industry Consortium, an organization that includes HP Enterprise, IBM, Oracle, and Quantum, along with a slew of academic research groups, released what it called the “International Magnetic Tape Storage Roadmap." That forecast predicted that the areal density of tape storage would reach 91 Gb per square inch by 2025. Extrapolating the trend suggests that it will surpass 200 Gb per square inch by 2028.

The authors of that road map each had an interest in the future of tape storage. But you needn't worry that they were being too optimistic. The laboratory experiments that my colleagues and I have recently carried out demonstrate that 200 Gb per square inch is perfectly possible. So the feasibility of keeping tape on the growth path it's had for at least another decade is, to my mind, well assured.

Indeed, tape may be one of the last information technologies to follow a Moore's Law–like scaling, maintaining that for the next decade, if not beyond. And that streak in turn will only increase the cost advantage of tape over hard drives and other storage technologies. So even though you may rarely see it outside of a black-and-white movie, magnetic tape, old as it is, will be here for years to come.

(Article is dated 28 Aug 2018)

 

[Sycraft]

Interesting example. I'd all but forgotten about magnetic tapes. I knew they weren't extinct, but didn't realize they were still actively developed.

It's niche to be sure, but still a big market. Some of it today is just obfuscated in that a company will outsource their data to a cloud provider, but use the ultra cheap tier storage... which is usually tape drives underlying it hence why it can take a day or two to get the data back. They have to fetch the tapes, load them, copy the data to something online, then get it to you.

Also something to be said for it with ransomware/other cyber attacks. Maybe an attacker gets in and owns every system, every server, and blows all that away. They aren't going to be able to do anything to the tapes sitting on a shelf.

 

[Lakados]

It's niche to be sure, but still a big market. Some of it today is just obfuscated in that a company will outsource their data to a cloud provider, but use the ultra cheap tier storage... which is usually tape drives underlying it hence why it can take a day or two to get the data back. They have to fetch the tapes, load them, copy the data to something online, then get it to you.

Also something to be said for it with ransomware/other cyber attacks. Maybe an attacker gets in and owns every system, every server, and blows all that away. They aren't going to be able to do anything to the tapes sitting on a shelf.

Every couple of years I call up Dell and order a fresh set of tapes. Accounting is upset on the amount of time it takes to back up to tape and they are getting annoyed but I just remind them how easy it is to damage a 2.5” mechanical drive and they give up the argument. Though I quoted them on one using SSD’s so who knows maybe they say yes? Telling them to simply generate less data doesn’t seem to get the reaction I want from them.
But we have a tape backups that happen M-F (each to their own tape), then one for the first Monday and second Friday of a Month then one that gets archived with the year end rollover.

 

[Sycraft]

Every couple of years I call up Dell and order a fresh set of tapes. Accounting is upset on the amount of time it takes to back up to tape and they are getting annoyed but I just remind them how easy it is to damage a 2.5” mechanical drive and they give up the argument. Though I quoted them on one using SSD’s so who knows maybe they say yes? Telling them to simply generate less data doesn’t seem to get the reaction I want from them.
But we have a tape backups that happen M-F (each to their own tape), then one for the first Monday and second Friday of a Month then one that gets archived with the year end rollover.

We do a full backup once a quarter, new tape set, and then differentials each week. Dailies (and hourlies) I let the NetApp handle. There's lots of repeated data with all the full backups, but that is the kind of retention they want and they want it kept for years. They want to be able to say "What was the status of this file in Q2 2018?" and this allows us to have that information, without having to pay for a half a petabyte of online storage.

Also, I don't worry about tapes on a shelf. Drives, those can seize up if you leave them sit long enough. Likewise SSDs need to be able to shuffle their data to new pages from time to time. But tape? Nah, you can leave that shit on the shelf for a decade, it'll still read (if you have a compatible drive).

 

[Lakados]

We do a full backup once a quarter, new tape set, and then differentials each week. Dailies (and hourlies) I let the NetApp handle. There's lots of repeated data with all the full backups, but that is the kind of retention they want and they want it kept for years. They want to be able to say "What was the status of this file in Q2 2018?" and this allows us to have that information, without having to pay for a half a petabyte of online storage.

Also, I don't worry about tapes on a shelf. Drives, those can seize up if you leave them sit long enough. Likewise SSDs need to be able to shuffle their data to new pages from time to time. But tape? Nah, you can leave that shit on the shelf for a decade, it'll still read (if you have a compatible drive).

Yeah I do like tape for the longevity, Pensions Canada requires us to keep our data for 99 years, so we have the bulk of it still on paper. Dell, HP, Ricoh, and everybody else can give me all the sales pitches they want for file digitizations. I barely trust them to support that tech after a decade let alone 10 of them. I only need the tape drives for recovery from malicious actors. We go with tape because you can drop it, we’ve had some butterfinger accounts in the past, most mechanical HDD’s don’t bounce.

 

[ND40oz]

Yeah I do like tape for the longevity, Pensions Canada requires us to keep our data for 99 years, so we have the bulk of it still on paper. Dell, HP, Ricoh, and everybody else can give me all the sales pitches they want for file digitizations. I barely trust them to support that tech after a decade let alone 10 of them. I only need the tape drives for recovery from malicious actors. We go with tape because you can drop it, we’ve had some butterfinger accounts in the past, most mechanical HDD’s don’t bounce.

You can’t drop AWS or Azure, object lock will keep it from getting overwritten or deleted.

 

[cjcox]

Broadcom to buy VMware in $61 billion acquisition.

"Oh, so you're talking about magnetic tapes."

 

[Fenris_Ulf]

The company I work for was part of EMC at one time. When EMC was acquired by Dell, Dell didn't know what to do with any of the software companies that were part of EMC (they just wanted the enterprise hardware and customer base), so decided that benign neglect was the best option and mostly left us alone since we just quietly sat back and made money. Besides the usual big corporation BS that trickled down, it was pretty OK. We always knew that Dell would eventually get rid of us for the right price, the question was when, not if. Dell just doesn't do enterprise software and doesn't want to be in that game. I also saw that VMWare was in the same boat, a sale waiting for the right conditions. $61B is a bit over-valued but 6-8 months ago when this deal started it probably seemed like a fair price.

 

[Lakados]

You can’t drop AWS or Azure, object lock will keep it from getting overwritten or deleted.

I can, but you can’t hack tape. AWS and Azure data storage gets breached all the time. The data is filled with a crap load of sensitive information. I trust the bad actors out there to find a way in more than my ability to secure it. That and we’d be uploading hundreds of GB a day, not sure my internet could handle that.

That and the software last got its overhaul in the early 2000’s and uses the best security practices available at that time. Which is to say basically nothing, pretty sure you could brute force it with a Cassio.

 

[lopoetve]

I'm surprised VMWare is even as big as they are anymore.

They were the only game in town in the beginning, but already back maybe ~8 years ago it felt like everyone, even big enterprise users, were in the process of transitioning away from VMWare to alternatives like KVM.

They all came back. None of the alternatives worked at that scale.

 

[cjcox]

They all came back. None of the alternatives worked at that scale.

Well, maybe if your annual bill for VMware is in the hundreds of millions (which isn't hard to imagine with VMware pricing). I'd say your statement is "false". Lots of business either went to cloud (which is bigger than big on virtualization and doesn't run on VMware) or moved to something KVM based or Xen based. While, my company did move from a community hypervisor to VMware, but not nearly to the same scale... well, because it's just that high priced (who cares about scale if it's unaffordable). Longer term, we're pushing everything into "the cloud".

 

[lopoetve]

Well, maybe if your annual bill for VMware is in the hundreds of millions (which isn't hard to imagine with VMware pricing). I'd say your statement is "false". Lots of business either went to cloud (which is bigger than big on virtualization and doesn't run on VMware) or moved to something KVM based or Xen based. While, my company did move from a community hypervisor to VMware, but not nearly to the same scale... well, because it's just that high priced (who cares about scale if it's unaffordable). Longer term, we're pushing everything into "the cloud".

Xen is effectively dead. KVM took some, but most ended up coming back given the extra hardware costs (no overcommit on a lot of stuff with KVM, even still - I last checked a year ago or so, granted).

Now the cloud? ABSOLUTELY. Hell yes they moved there - but on-prem alternatives? Edu did hyper-V due to licensing benefits, and some oracle went to OEL, but a significant number of the medium+ sized shops ended up coming back after failed or mediocre experiences with KVM (even in Acropolis form). It just didn't quite have the feature sets, and no one wants to try that migration twice if it didn't go great the first time. :-/

 

[cjcox]

Xen is effectively dead. KVM took some, but most ended up coming back given the extra hardware costs (no overcommit on a lot of stuff with KVM, even still - I last checked a year ago or so, granted).

Now the cloud? ABSOLUTELY. Hell yes they moved there - but on-prem alternatives? Edu did hyper-V due to licensing benefits, and some oracle went to OEL, but a significant number of the medium+ sized shops ended up coming back after failed or mediocre experiences with KVM (even in Acropolis form). It just didn't quite have the feature sets, and no one wants to try that migration twice if it didn't go great the first time. :-/

We ran oVirt and I could totally do more with it and faster than anything VMware has. VMware is old. It hasn't changed all that much from 10 years ago (or more). Just saying.

There's a ton of advantages in not choosing VMware. I think if you really talk to people that have truly used something else, and by that, something that was done well (like what you'd have with VMware), most would tell you it's night and day the advantage of not using VMware.

If money is no object and you've got time to deal with typical VMware issues... sure. I suppose.

 

[Lakados]

Well, maybe if your annual bill for VMware is in the hundreds of millions (which isn't hard to imagine with VMware pricing). I'd say your statement is "false". Lots of business either went to cloud (which is bigger than big on virtualization and doesn't run on VMware) or moved to something KVM based or Xen based. While, my company did move from a community hypervisor to VMware, but not nearly to the same scale... well, because it's just that high priced (who cares about scale if it's unaffordable). Longer term, we're pushing everything into "the cloud".

I’ve looked at KVM, Scale, HyperV, etc there is lots that they don’t do that VMWare does. Even things as simple as VMWare essentials has features that aren’t present in the others or aren’t as seamless.

And I do have a bunch of HyperV stuff but we’re likely going to consolidate and just go back to VMWare across the board, it simplifies the airgapped backups.

 

[lopoetve]

We ran oVirt and I could totally do more with it and faster than anything VMware has. VMware is old. It hasn't changed all that much from 10 years ago (or more). Just saying.

There's a ton of advantages in not choosing VMware. I think if you really talk to people that have truly used something else, and by that, something that was done well (like what you'd have with VMware), most would tell you it's night and day the advantage of not using VMware.

If money is no object and you've got time to deal with typical VMware issues... sure. I suppose.

Disclaimer: I worked there as a senior engineer and have contributions to the code. I also have worked for one of the largest OEMs directly supporting KVM for a while, and now for a systems integrator supporting all of them. If you think VMware hasn't changed in 10 years (2012 was 5.X timeframe) to now... I'm curious how much you've used outside of the free tier, and how modern the equipment was?

Suffice to say, I'd love to see examples of what oVirt could do that VMware can't, because I've got a massive list of ones VMware can that oVirt can't (like overcommitting CPU cores, for a basic one, or truly automated resource management, or better-than-best-effort HA, etc etc etc).

I’ve looked at KVM, Scale, HyperV, etc there is lots that they don’t do that VMWare does. Even things as simple as VMWare essentials has features that aren’t present in the others or aren’t as seamless.

And I do have a bunch of HyperV stuff but we’re likely going to consolidate and just go back to VMWare across the board, it simplifies the airgapped backups.

Yep.

 

[pillagenburn]

They still make the best, most complete, hypervisor solution. We use Hyper-V here, licensing costs got us to change mostly, but VMWare really is a cut above. I still use it for desktop stuff and play with the enterprise stuff occasionally and particularly if you want to do a mixed OS environment it just kills. It is expensive, but for some the cost isn't an issue, it is worth it.



Yes and no. While plenty of businesses went wild shoveling everything out to the cloud as fast as they could, others have not. Also some of those that did are discovering a few things about the cloud:

1) The cloud is only cheap if you don't use much. There's a reason why MS and Amazon have such stupidly large profit margins: They charge the fuck out of you for cloud service. If you need a system full time, it can easily run you the complete cost of a server in 2-4 months to have it running. So it can be cheap if you have a workload that scales up and down real fast and you can keep it lean and mean, but that isn't the case for a lot of things. While paying monthly masks the true costs against people who don't know how to annualize things, eventually there is some executive smart enough to do that and say "We are paying HOW MUCH for this?"

2) Latency/availability can be a real issue. Shovel everything out to another data center, and suddenly there is latency that wasn't there before. This can be an issue for some things. So you start to see some silly things like "edge cloud" which means "Put a server on prem, that you pay for, but pay a cloud provider to manage it."

3) Some people are wondering if maybe giving all your data to Amazon, a company not known for being super ethical, is perhaps not the best idea. Sure they claim they won't be looking at your shit for trade secrets, but you have only their word on that.


At any rate, some companies are not as interested in the cloud, or only want it as part of their solution. For them, VMWare may be an answer. They have some slick stuff if you want to do things like VDI where you build up an on prem solution that meets your normal load, but have cloud availability to deal with peak load. That way you aren't playing out the ass for stuff you are using all the time, but if you don't have to have a bunch of extra hardware sitting idle 99% of the time to meet the occasional demand spike.


Always remember that just because there's a new, shiny, trendy solution it doesn't mean the old solutions necessarily go away. One of the examples I love is tape storage. Not only is it still used, it is still under active development. We just bought a brand new LTO-9 tape drive. Sounds archaic, and there are things about tape that suck... but it is $5/TB for data storage and that's a one-time cost and scales basically infinitely. You can packrat data at a huge scale for a price way less than you are going to get with newer technologies.

Speaking of mining for trade secrets - any idea if other cloud-based services do the same or know of specific instances where services have been caught doing this? Salesforce comes to mind.

 

[Lakados]

Speaking of mining for trade secrets - any idea if other cloud-based services do the same or know of specific instances where services have been caught doing this? Salesforce comes to mind.

Google got caught doing it, or more to say a single google engineer got caught some time back snooping in on accounts looking for things to fill their spank bank.
Amazon and Microsoft encrypt your instance and they don't have the keys, if you lose your credentials it is actually a very unpleasant process to get it recovered, which is why they recommend having a break glass account that you have the information written down in a safe location just in case.
But as Amazon proves time and time again, if you set your account up wrong you leave it open to everybody and their dog, so many unencrypted data stores out there on Amazon, so who knows if they are snooping on their own unencrypted ones. But Microsoft does not let you create an unencrypted data store.