After 114 days of change, Broadcom CEO acknowledges VMware-related "unease" | Hacker News Reader

After 114 days of change, Broadcom CEO acknowledges VMware-related "unease"

arstechnica.com

11 pointsRyeCombinator16 comments

You know there are few technologies that I like, take C#, it's an awesome language, but satan forbid investing efforts to make myself subjects to a change of CEO, an acquisition, anything, one thing I take from all of this Broadcome situation: always use only open source

C# is open source

Its very dependent on a corporation, maybe I could rephrase it as only choose software not tightly coupled with a single corp

Love the idea. I'm not sure if anyone has found a sustainable model yet.

Looking at the current CNCF landscape, majority of the graduated projects have huge backing from single corps - promoted with the open source labelling.

https://www.cncf.io/projects/

i.e. argo by Intuit, Cilium by Cisco, SPIFFE/SPIRE by VMware/HPE

And this is why I keep coming back to Common Lisp

Underrated comment

The fact that there are many projects depending on a single corp doesn't mean there aren't projects who are supported by community or a group of corp, I don't know take rust, take podman

It's unlikely to ever happen that MSFT just says "we're not maintaining this anymore, we're making our own private fork". This seems like a huge blow to reputation and no financial upside. Kind of like Google abandoning Go (but then again, it's not like Google does not have a reputation for killing products).

The main conundrum is the sheer cost of maintaining dotnet and its surrounding infrastructure, and the teams to keep it moving forward fast.

But should unthinkable happen, it is MIT licensed so one or more companies can just carry on with their fork or maybe even transfer existing org/repos to some newly formed foundation.

I would use Go more confidently than C#, C# depends on Microsoft for everything, from the tooling to asp.net, entity framework, c# is mit licensed but you can’t get anywhere without microsoft.

Now I understand there's little chance microsoft kills it, but I would assume people would have thought that of their perpetual licenses or plans for VMWare, they maybe thought they're not going to kill their clients, they're not going to kill sales partners because XYY, the truth is, there are good technologies very dependent on corporations deciding their directions, but those technologies aren't that level better than open ones to justify the risk

Who develops Go? :)

Yeah I understood that, but I said "I would rather" , not "I will" , I think the ecosystem of Golang is much less dependent on Google than C#, there are a lot of tools in Go that aren't related to Google and in case they'd give up the language, it would be much easier for the community to keep maintaining it, I think

Unity, games written with Monogame or Godot, frameworks for distributed actors like Akka.NET, all kinds of message buses, AWS lambda for serverless that works on top of NativeAOT, Dapper and DapperAOT, AvaloniaUI (while people bash .NET for GUI story, they don’t bash Go because it just doesn’t have any good options at all) and Uno platform.

No, it doesn’t seem to be the case. And C# mostly does not need multiple back-end frameworks because the out of box option (which is just a part of dotnet foundation ecosystem of repositories) provides the best performance out of all alternatives like Servicestack. This does not mean someone won’t implement another one in the future to sacrifice even more ASP.NET Core style features to minimize the overhead even further or to address specialized scenarios.

Oh yeah probably gaming in c# is a industry little dependent on Microsoft

if you're still using vmware for virtualization i feel like you get what you deserve at this point. KVM has been around for over a decade, powers most clouds, and has easy to use and powerful implementations.

VMWare by comparison is hot garbage. Want to migrate a VM? thats a fee. Want to clone a VM? thats a fee. need a way to configure more than one hypervisors networks at a time? thats a fee. want iSCSI? you've gotta dedicate a full network port to the traffic and cant converge it with the existing trunk. need ARM? tough shit. want support? good luck, if youre calling VMWare you are truly desperate, and some of those 2800 people cut from the rank and file probably had byzantine ESX style answers for you.

VMware is more than a hypervisor.

At $WORK, we use VMware - and the "hypervisor" part of it is honestly not even close the selling point for anyone doing serious enterprise workloads - and VMware knows this.

Smaller firms with a few VMs here and here could probably lift and shift into another platform super easily.

For the bigger firms with large on-prem investments is where the vendor lock-in is the most difficult. Take two parts of VMWare:

- NSX-T ([1]): The SDN and networking overlay component of VMware. It provides overlay networking and micro-segmentation (like your AWS VPCs, etc.), as well as service insertion for firewalls, etc. Pretty much nothing competes in this market for on-prem that's actually integrated with the hypervisor.

- vRA ([2]): Does all your infrastructure automation, monitoring, and optimization. Again, it's actually integrated with the hypervisor.

Not at all part of the hypervisor, but components that are part of virtualization infrastructure you just can't go without.

Now, I'm sure you could argue that there's other systems you could fit in here, or tooling that exists that solves these problems - won't argue with that. But the big part is integration - there's basically nothing that you can buy off the shelf that does all of this under one roof that doesn't become some cobbled mess of poorly-integrated tools.

Is VMware perfect? By no means. But at the end of the day, most companies want 1) something that works, and 2) someone to call if it breaks. Broadcom knows this - hence why they placed a bet that companies too deep into it couldn't shift off quickly enough - and I fear they've betted correctly.

The closest would maybe be OpenStack? Who knows where things will end up in the next few years!

[1] https://docs.vmware.com/en/VMware-NSX/index.html [2] https://docs.vmware.com/en/vRealize-Automation/index.html

Reply on news.ycombinator.com