Hacker News Reader

Comment by nerdjon | Hacker News Reader

First, question:

> Several top VMware executives have left the cloud-computing company over the past year.

Since when is VMWare a "cloud-computing" company?

I am now reaching out to friends to see if everyone is ok, this sucks. I don't really get doing layoffs as soon as the acquisition finishes. Wouldn't you first expect that people are going to leave just due to them being bought and then do additional cuts if necessary after that?

I feel like that has always been the strategy with acquisitions minus redundancy like HR.

Otherwise this is a big moral hit, first being bought and then an acquisition at the same time? Who would stay after that?

View on HN

> Since when is VMWare a "cloud-computing" company?

I think most VPS hosting providers are running on VMWare. Whether that counts as "the cloud" or not is up to you, but virtualization within datacenter is pretty standard.

For what it's worth, every experience I've ever had with Broadcom in the embedded space has been negative. So this came initially as a disappointment, and now I'm disheartened to see my expectations were accurate.

> I think most VPS hosting providers are running on VMWare.

This is highly unlikely. VMware is too expensive for a VPS provider to achieve profitability.

It seems highly unlikely they're not successful and selling for $64 billion. I don't get it either :/

F500 companies are paying them millions in licenses to save 10x their hardware cost.

It's usually proxmox or virtuozzo, or some other kvm front end. I've worked at 4 huge webhosts and they never used vmware for vps/shared servers. Only our dedicated customers paying for support (think RAX) used vmware and that was their own.

VMware is extremely common in hosted solution providers. This is not just web hosting, this is hosting stuff like your payroll or ERP systems.

There a big difference between hosted solutions and VPS services. VMware might be used a lot in the former, but in the latter it's mostly KVM or Xen.

I don't know if it's 'most' but there's plenty of large private-cloud B2B setups offering vm-as-a-service / esxi-as-a-service or something frontended by vmware cloud director. Definitely to the tune of billions of annual revenue worldwide.

I worked at some fairly large hosting companies before cloud took over and never once saw vmware. It was almost always virtuozzo/openvz

There are two classes of hosting companies, enterprise and small business. Enterprise hosting companies serve enterprise fortune500 customers. They are VMWare. The ones catering to small business are using everything but VMware because of costs. They are not typically large, digital ocean and linode aside.

I still disagree. Verizon serves every fortune 500 company in existence and our enterprise compute offering was not vmware. Not a single vmware node in all 100+ pops. I also worked at Godaddy, zero vmware, even after IPO, and they most certainly have B2B customers too.

I have never used a non VMware environment at a Fortune 500.

I worked at one of the biggest private cloud providers in Australia.

It was VMWare VCentre/VSphere end to end for all private clients.

There were some great GUIs and great management but if you wanted serious API access it was the one of the most hacked up thing I have ever seen in my career.

Oh my lord, I remember dealing with a VSphere API early in my career. They didn't have a concept of service accounts so the official advice for our nightly automation was:

Pay a sysadmin to be up at night, then before your job is meant to run, have them log into the VSphere GUI, open devtools, click refresh to trigger a GET request, copy the cookie, paste the cookie into our custom software, and send it in the cookies of our API requests.

When we asked if there was an easier way, they said that we could probably use Selenium to automate that.

Needless to say, we did not sign the contract, and moved wholesale into AWS.

That's about it. Although, instead of using Selinium, you could work out the XML that the GUI runs from and `POSTS`, then make scripts pretend to be the GUI. Then maybe going into the internal postgres database that runs it to get the data that is not on the GUI.

Not funny. :)

Oh haha I forgot how much information was not exposed in the API at all. I remember having to bring in a senior engineer because I thought I was going crazy. I could only ever find like half of the useful data on a server I just provisioned.

As someone with a vast amount of experience automating vsphere, this is made up bullshit.

API access got a lot better once they introduced the powershell API (which can also be used via C#) but the solution you're describing is the result of people not knowing what they're doing.

It looks like the automation api was introduced in 2021. (!!!)

I stopped using it in like 2014 or 2015 after years of requesting improvements (like better ways to automate things without stuff breaking all the time).

Me too. I have not touched it for over 2 years and big clients sre unlikely to make huge upgrades quickly anyway.

It sounds great they have arrived at the 20th century. Powershell and C#? Don't tell me it is SOAP and all XML? I kinda like the way most cloud apis are just REST and json under the hood with vendor supported python packages that work well on all major platforms.

This sounds like a Ruby on Rails developer who cannot fathom that not everyone thinks having to update their framework every 6 months is unacceptable.

PowerCLI is something like 10+ years old, and all it did was used the existing API's but put a pretty sheen on it for those that were confused by the graph ideas used in their full-on API.

This was almost a decade ago, I assume (hope?) things have changed in a decade of trying. But it was official VMWare advice! It was literally in their documentation too.

In my experience in these situations (I was always being acquired), the acquiring company is at least partially counting on cost saving from getting rid of departments like finance, marketing, sales and legal as they believe they can do that with their own teams. The acquiring company also will have their own metrics for proper sizing (employee count and spend) for engineering, support, product, etc which is likely lower than the company being acquired. But in this case VMWare is pretty mature so those metrics maybe ok Can’t say I support this, but the bean counters love it.

Edit: “ Broadcom has not returned CRN’s request for comment about the job cuts. The company, which closed its acquisition of VMware Wednesday, had previously stated it could find $250 million in synergies once it completed the deal.” Synergies is code for the redundancies I mentioned above.

    Synergies

Also: Less competition.

> I think most VPS hosting providers are running on VMWare.

Out of date info, but last time I really looked there used to be 3 tiers of VPS provider:

* VSphere with SAN and blades (I think there was an HP reference implementation for this - many providers seemed to use it). These were by far the most expensive, and not really very good performance for the money due to SAN speeds and ESX not cpu scheduling multi core VMs that well.

* Xen/KVM. Mid range in terms of cost, and less fault tolerant than the SAN backed VSphere stuff. But usually out performed it.

* VServer / OpenVZ at the very cheap end. These were more like early LXC containers than VMs. The price point meant that you got what you paid for, and there were tech limitations with what you could do with the kernel (they were sharing the host kernel). Security wasn't great either.

I doubt the VMWare providers are as common now as they were 15yrs ago though. Even back then, they weren't the most common though.

VPSes use Xen and such, not VMware because of $. VMware was born out of the need to increase utilization of expensive enterprise servers by putting many virtual machine guests on fewer physical hosts.

>I don't really get doing layoffs as soon as the acquisition finishes.

Planning layoffs is typically part of the M&A process, so when it completes, they typically already know who is going to be let go and drop the axe at that point.

The people making the cutting decisions are also a removed from individual IC level visibility. These cuts are often done based on financial metrics per product line.

You could be an amazing engineer and have a well engineered product, but still get laid off because the product didn't get market traction.

I've been on both sides of the equation. It sucks but such is life sadly.

And ability doesn't fit in the spreadsheet they use to calculate their target profitability.

They are going to cut expenses by 40% and increase costs of their products by 20%.

That is the point of lay offs. They are not saying you are bad they are saying they don't need someone in your position.

sometimes a company will let managers in a different department that is hiring know about coming lay offs and ask/force them to take good people. this is rare though.

> sometimes a company will let managers in a different department that is hiring know about coming lay offs and ask/force them to take good people. this is rare though.

Depends on the kind of layoff. If it's strategic M&A related this happens fairly often. If it's purely a cost cutting hail mary, then not as much.

There's probably some legal loophole by which _technically_ the employees have a contract with a new company after the merge, since their old employer no longer exists on paper.

This means employees get let go with legal bare minimum compensation and have little (if any) legal position to fight back.

Immoral but not illegal.

In the U.S. (where VMware is based), virtually all jobs can fire you for no reason at will. Contractually guaranteed jobs are rare.

> This means employees get let go with legal bare minimum compensation and have little (if any) legal position to fight back.

I don't know what country you're posting this from, but in the US none of this matters, and it does not apply. When someone is let go there is no "legal bare minimum compensation." There are some notice requirements (e.g. the WARN act), and many companies will pay out compensation in lieu of notice if the WARN act applies, but an acquisition makes no difference to these notice requirements.

In the UK we do have some protections depending on length of time with the company. There is some framework for redundancy payments.

The US culture of "at will employment" (i.e fire you on the spot for little reason) gives me anxiety and I don't even live there!

There are pros and cons. The ability to lay someone off quickly means someone in the US can risk hiring a bunch of people on a risky bet - if/when they give up on the project they can let the people go and cut their losses. Some of those risky bets pay off though and the people are not laid off.

Note that I said lay off not fire above. There is an important legal difference.

There are both side of the coin when it comes to at-will employments with layoff. They can use it to get rid of high paying employees to hire lowest quality of employee for little pay that will affects the quality of their product. And they don't care because it looks good on their balance sheet. And there is no recourse for high paying employee against the company (unless they have a signed contracts that have a clause to prevent this).

My former job laidoff few people because of high paying salary. They got high salary because positions required a specialized and niche knowledge for their jobs. They are good at what they do which earned their justification for high paying. It took the company a year to hire a barely qualified person because they don't want to increase the pay after declined job offers from multiple potential candidates.

> an important legal difference.

so what is the difference?

a layoff is due to business "loss" or changes? How is that really any different to a firing?

Fire is for cause and that means they will tell others when asked. This also means more paperwork to prove you are bad. That is they want someone doing the job you are supposed to do,but for whatever reason not you.

Laid off just means they don't need the job done anymore, and has nothing to do with how good or bad you are.

> Laid off just means they don't need the job done anymore, and has nothing to do with how good or bad you are.

Eh, in theory maybe. Not really in practice. When layoffs happen, companies often prioritize laying off lower performers. And they often still need the jobs done, just fewer people.

Of course none of these are legal differences. The legal difference I know of is that being fired for cause affects whether you are eligible for unemployment insurance benefits.

Sure is nice when you hire a bad employee

sure is nice when you have a bad employer

Then find an alternate employer?

The whole problem is employers abusing employees by firing without any reason, and using the threat of firing employees if they complain. That is used to force employees to accept unsafe and illegal working conditions, because if you're an employee who's choice is "report workplace problems" or "pay rent" you're kind of forced into accepting unsafe or unreasonable work places.[1]

Because it turns out that is what employers do. That's why most countries have have laws preventing employers from arbitrarily dismissing employees unless you have an actual reason.

If you have a bad employee, you can fire them. You don't need at will employment for that. Unless you mean "I don't like this employee, who is doing their job properly" in which case you don't have a bad employee, you're an asshole.

[1] If you're ever fired and your employer claims they're allowed to do so because "at will" employer, note that "at will" only covers "legal" reasons. Retaliation, bias, anti-pregnancy, anti-child, anti-religion, etc are all illegal reasons. Always ensure that everything you communicate with your employer is in writing. As @nxm is demonstrating your only safe option is to assume that your employer is your enemy.

In Germany, we have a 6 month "probationary period" during which either side can terminate the employment contract. That's more than enough time to check if you got a dud.

Especially notable in an easy to fire environment when health insurance is tied to employment. Everything feels higher stakes when there is minimal social safety net.

Ever since the ACA passed, you can jump right onto a marketplace plan with very little out of pocket (based on your income, which is $0 if you’re unemployed!). And they can’t exclude preexisting conditions anymore. Would be even better if employer-based health insurance went away completely, but at least it’s better than the old status quo.

To clarify my earlier comment I know "to fire" and "to let go" are quite different - the former implies you broke contract, be it illegal means or poor performance reasons.

"to let go" means redundancy.

It appears some countries see little difference between the two and others have a clear distinction and legal process.

Why is it immoral to fire employees that the owner thinks are not needed?

If Broadcom lets the VMware employees decide to leave on their own, they will be losing the top performers and those with highest prospects of getting a job elsewhere quickly. They end up left with the less good employees.

But if instead Broadcom proactively sheds the underperformers, then the desired employees are reminded to be scared of unemployment by the carnage. And at the same time the market is flooded with a bunch of people looking for the same jobs a high performing employee would have, so now it's harder to jump ship.

All this is disastrous for morale, but no worries, the beatings will continue until morale improves.

Back in 2020, VMware pivoted hard into the cloud computing/software space with the acquisitions of Carbon Black, Heptio, and Pivotal—I was a Pivot prior the acquisition. They rolled all these acquisitions into a single business unit to sell (formerly) Pivotal Cloud Foundry (PCF), Pivotal Container Service (PKS), and other products in an on-prem, enterprise PaaS offering.

I left shortly after the Pivotal acquisition closed, so I didn't see firsthand how it unfolded, but from folks I talked to, it seemed that VMware batted teams around and really mixed the whole thing up.

Picked up where you left off. It seems that VMware was pulled apart and its carcass was sold off

>" They rolled all these acquisitions into a single business unit to sell (formerly) Pivotal Cloud Foundry (PCF), Pivotal Container Service (PKS), and other products in an on-prem, enterprise PaaS offering."

So what happened to Heptio exactlY? Were they combined with Pivotal and spun-off?

Their staff got smooshed together with Pivotal and a few other acquired odds-and-ends and largely set to work building a suite of Kubernetes-related products under the “Tanzu” brand. This included a Kubernetes manager (TKG) and an application platform (TAP) that were supposed to replace the BOSH-based Kubernetes installer (TKG-i, formerly known as PKS) and application platform (TAS, formerly known as PCF, better known as Cloud Foundry).

They got stirred around a few times via reorg, then the Heptio founders retired. Last I heard everyone in Tanzu was trying to get themselves reclassified as somehow part of the Cloud Foundry/BOSH stuff because it has significant revenue and was believed to be safe(r) from Broadcom layoffs.

Wow, what a mess. Thanks for the recap. So is the Tanzu product going away then and were they hit hard in these layoffs?

The exact opposite. Tanzu is a cash cow due to their customer base from pivotal. Its a mix between they left the good parts of the Tanzu portfolio alone while they chased their Kubernetes vision. The commercialized version of Cloud Foundry never went away post pivotal acquisition, the “legacy” products VMware tried so hard to move away from accounted for a overwhelming majority of Tanzu’s revenue stream

disclaimer: ex tanzu support

Yeah I don’t know what they’re doing with the new parts of Tanzu that don’t have revenue (will ask the next time I talk to folks, but I mostly don’t know people in that part of the org) but I haven’t seen any sign they’re cutting TAS. The only folks I know who got laid off so far took an optional severance package that was offered to everyone. The continuing offers were strong.

Cloud Foundry customers really, really like Cloud Foundry, and are uninterested in running it without support from a vendor.

vmware is very popular in the private IAAS space. If you have special legal requirements but are too small a fish for the big players to spend time catering to, this is where you end up. IAAS operators that are competent at operations and business logic but not necessarily having the engineering talent needed to run a kvm fleet.

> Since when is VMWare a "cloud-computing" company?

That's like 90% of what they do now.

They actually make that really clear with little cloud icons on the "Product" menu on their homepage.

> I don't really get doing layoffs as soon as the acquisition finishes.

It is pretty standard procedure. Buyer usually absorbs leadership first, then they gut sales, then marketing after that. Leadership is kicked out or leaves after a while.

They are treated like redundancies, even if they aren't so.

I feel for the VMWare folks. I may have my issues with their pricing, but their products have historically been quite solid, while Broadcom strategy is, well, just "buy and spread".

VMWare is used in the cloud as an alternative option from docker container.

VMWare is also diving into K8S space, I'm sure they have some sort of "managed" service offering that bundles their K8S solution.

VMWare owns Spring Framework... that's a cloud toolkit.

Vmware bought heptio and with that things like their managed control plane (a la Google’s Anthos) and other k8s stuff. I think most of those folks are gone now though

The headline folks from Heptio are gone but the products and staff are still there. They’ve gotten a lot of internal transfers from other parts of VMware.

Thanks for pointing out the Spring connection... I wonder how that will impact their development.

https://x.com/odrotbohm/status/1729231722498425092?s=20

If it's true that they'll lay off Oliver Drotbohm and Dan Vega then "seriously bad" could describe the impact.

> Since when is VMWare a "cloud-computing" company?

Well they have this Pivotal Cloud / Cloud Foundry /VMware Tanzu crap. I know this isn't what people mean when they talk about "cloud company" but it works as "private cloud" in enterprises.

>crap

What do you suggest as a replacement to CF and Tanzu?

We have had sales people trying to push this to us for the last few years and so far it's been ridiculously priced and nobody seems to come up with plausible real life success stories.

My impression is that vmware have had almost no traction for any product beyond stock vsphere and even there the API's see very little use as most customers prefer to point and click in the web interface and are basically happy with what vmware delivered a decade ago in terms of features, but people in this market are also loyal to vmware as it's the known way to get virtualization to smaller server rooms.

Our Full hosting service / IaaS / PaaS builts on ESXi / vSphere HA, with custom or OTS orchestrators. It is quite common actually in service provider clouds, at least the ones that are more traditional and not relying on fancy, nwer stacks.

They sell the tools to build private clouds.

Reply on news.ycombinator.com