Hacker News new | comments | ask | show | jobs | submit login

Guys, this is not just, or mainly not RH Linux. :(

- kernel development

- Ansible

- JBoss (I know HN hates Java, especially Java EE, but it was and is an important factor in enterprise OSS adoption)

- OpenShift

- Ceph, Gluster

All these are in danger, not just RHEL. I don't know about any other company that is large, successful, focuses on the enterprise and absolutely behind OSS. Canonical is way behind Red Hat in terms of revenue (1/20).


Also in practice glibc, systemd, GNOME, etc. have a ton of maintainers from Red Hat, even if they're not Red Hat-owned the way Ansible or Ceph is.

Yeah the whole freedesktop.org stack is to a large extent RedHat folks.

Not to mention LibreOffice.

I don't know of any other OSS company that does _exclusively_ OSS.

Canonical, for example, has a number of proprietary solutions built around their core OSS stuff, so they function as an open core business. And they're not doing anywhere close to as well as Red Hat does.

WSO2 - an open source integration software company - is exclusively OSS. We will do $50M in sales this year with 80% of that subscriptions for on-premises open source software. We are the 6th largest OSS company by revenues. We'll probably jump up to 5th next year because of the RedHat acquisition of IBM.

I'm happy for you, really. But you being in 6th place with 50M$ in revenue says worries me about the market.

big typo on your home page "by using our our open source"

Thanks - forwarded onto marketing.

Nice to see WSO2 represented here!


SUSE? They make a big song and dance (literally) of being open source, but I don't know if they're exclusively OSS.

[I work for SUSE.]

Nowadays, yes. SUSE Studio was the main proprietary thing we had in recent memory, and it was sunlit a few years ago with KIWI being its successor.

As far as I know, everything we develop is free software. You can get the full sources for any package in SLES or openSUSE (which isn't really SUSE but SUSE engineers work a lot on openSUSE) using zypper.

Personally, not only is everything I work on free software, I also exclusively work upstream-first (and I maintain several upstream projects like runc and some OCI projects). To be clear, this is not a company-wide thing -- many of my colleagues do not consistently contribute upstream -- but regardless all of our products' sources can be downloaded under free licenses. In fact you can get the sources from OBS (it's what openSUSE Leap is directly based on).

Now, there are some things that we distribute which are proprietary to certain customers (think flash or the NVIDIA drivers), but these are mostly because customers pay us to repackage other peoples proprietary code. We don't develop them. Personally I'd prefer if we didn't do this, but it is a very small part of our business.

@cyphar - I would love to chat with you about SUSE's policies and encouragement to enable working exclusively on upstream-first. We are working to bake that concept into our policies at WSO2 (I'm its CEO). While we practice that in motion, we are codifying it within policies. If you could email me tyler@wso2.com, would appreciate a chat about it with your or management there.

Is openshift really open source? I was under the impression that nobody really runs origin on its own since it's far too hard. It seemed more like they just wanted to point out but it's open source, but you need support and a lot of know-how to actually use it.

OpenShift comes in two variants: Enterprise and Origin. Everything you can do in Enterprise can be done in Origin (Origin is ahead of Enterprise), but Enterprise comes with support.

I run Origin for my employer. It's pretty easy these days since you can do all of it with openshift-ansible and other related tools.

I know there are plenty of others like that, as well.

We run OKD (formerly Origin, the project was rebranded) as well and have for two years in just a couple weeks. It's been a pivotal part of our application stack since we adopted it, the developers on my team love being able to paste in a git URL and have an application live in a few minutes.

I stand corrected. Thanks.

We also run Origin.

we run okd as well...

I help maintain a Chef recipe that installs both - it's quite feasible but not for the casual user.



Add Fedora to the list - while technically a community project, Red Hat is the main contributor.

It seems to me that Red Hat is based on Fedora. Fedora being the fast release gratis community version and Red Hat the slow release commercial corporate version.

Then CentOS would the community gratis version of Red Hat.

That's basically how Red Hat describes it:

Fedora - Community Innovation

RHEL - Commercial Production

CentOS - Community Adoption

Also with RH's recent acquisition of CoreOS, that suite will also be in IBM's arena. I work for "a corporate". We used to be big IBM shops (AIX, IBM Java, websphere, etc, etc...). But AIX is now hardly used and neither is websphere. It's mostly off the shelf k8s, wildfly and other open source projects deployed on RHEL (not my personal favorite since its kernel is way behind the times). I think this is IBMs way of playing catch up in the arena. I worry how they will handle entrenched "political interests" - e.g. Webpshere/AIX/etc.. and the more open source friendly RedHat suite. Pretty sure its going to be one big mess in a few years time.

They _did_ pay a lot of money to acquire them, so there's a reason to believe they're committed to fully integrating RH technology. Considering Jim, who certainly has an interest in the survival of open source software, signed this off also gives some hope.

Just because the people in charge sign off doesn’t mean they won’t later be fucked over anyway. E.g. Instagram & Whatsapp.

The OpenShift folks sit on a ton of critical posts in the Kubernetes project. I'd suspect this is what IBM wants, but it's a risk that they stop contributing to the open source project.

It goes much deeper than that, Red Hat maintains or contributes regularly to a wide swath of the Linux ecosystem. They also have a policy of open sourcing acquisitions. If the deal goes through, it's a dark day for open source software.

- kernel development

- OpenShift

I very much doubt these 2 are in any danger. In fact, I'd say these are the golden eggs IBM wants.

Agreed. Seems that most $largeco I come across used to use IBM AIX and are switching to RHEL. This seems more strategic than surface level only.

You might want to read up on your Aesop’s fables ;)

Red Hat's Open Source only approach does mean that none of these projects can be closed down, though: AFAIK, IBM are acquiring people and a brand, but no significant proprietary IP. If higher-ups at IBM turn out to be as clueless as Oracle, then key engineers will just walk and continue working on those projects elsewhere.

Where elsewhere? The landscape of options to be paid for your OSS has just become a barren field.

Yes, OSS as a concept can survive only on gratis work. However I'm not sure the portfolio of projects supported largely by Red Hat maintainers could. If maintainers are forced to start walking as they did with Oracle I expect to see quite a few projects fall into disrepair.

I guess they can start a new company and keep working on the same products. The community would have to change to their new repo's but other than that RedHat could get a new start, with a different name.


Trademark is owned by Red Hat


Also Atomic/CoreOS - this will probably be cut now, perhaps even before it is complete.


I doubt it. When you read the IBM and Core OS acquisition press releases side by side you get a lot of the same "hybrid cloud" language.

Update: Near as I can tell the acquisition completed in March/April.

We'll see. Often talk like that is just buzzword soup. What matters to large companies like IBM is money from services, not owning the coolest tech stack. If it is clearly contributing lots of money or it has a champion in IBM they'll keep it.

In progress greenfield projects with no obvious monetisation are just the sort of thing that gets cut in this sort of merger, after the assurances that redhat will be run as a completely independent unit are forgotten and a new manager comes in looking to trim fat.

I use coreos and am now very concerned about its future.

Well over in Fedora, it is full steam ahead. CoreOS is going to be merged with "atomic" (in particular rpm-ostree style updates and layering), and then base Silverblue on that, which will become Workstation. Could all of that work just stall out? shrug Yeah sure, and Red Hat could suddenly decide they want to support Btrfs.

CoreOS is going to be merged with "atomic"

I’ve been following that work with interest.

Red Hat could suddenly decide they want to support Btrfs

Red Hat won’t be deciding much about anything after the merger goes through. I sincerely hope you’re right though and it survives and makes money somehow.

Pivotal is probably next in line for size for independent enterprise OSS but still a fraction the size of Red Hat, at under $1b in revenue. And some proprietary bits.

Also CoreOS, recently acquired by Redhat.

IBM would rather push Spectrum Scale/GPFS over Ceph or Gluster

Spectrum Scale Support is one of the most horrific things I've encountered in my entire professional career.

- etcd

Also everything CoreOS-related.

CoreOS :(

Super pessimistic point of view. Why do you think these are all in danger? These are things that people and more importantly enterprises use. Why would IBM ruin that?

Because IBM ruins everything it touches?

It’s possible they don’t want to. But they will.

Exactly, these are products to monetize and make money from.

Sure rationally the products should be a success, that's why IBM bought them. But GloboMegaCorps have a consistent record of acquiring great products then running them into the ground by smothering them with shitty internal policies.

It seems like a fair % of SVers haven't had the pleasure of working for one of these Kafkaesque giants. They operate on dream logic and risk aversion.

Though I dedicated a few years of my life to OpenStack at HP, I sometimes wonder if OpenStack would be healthier if GloboMegaCorps didn't get involved.

In 2013, OpenStack felt like the "new Linux"

2013, OpenStack at HP, aka Helion? Anupriya Ramraj - is that you? :) If so, your old team misses that project. We're bored these days.

> I don't know about any other company that is large, successful, focuses on the enterprise and absolutely behind OSS.

Doesnt Microsoft tick all those boxes? Even if they aren't exclusively an open source company, they are "absolutely behind OSS" if you go by how much open source code they have contributed.

How on earth is Microsoft "absolutely" behind OSS? Which of their main products is Open Source? Windows? Office? SQL Server? Azure? Exchange? What exactly makes Microsoft a company "absolutely behind OSS?"

Their stupid boot loader still ignores any other operating system for god's sake.

Kudos to people at Microsoft's Marketing and "developer relations" department who won the hearts of developers by allowing TypeScript and VSCode to be FOSS. Suddenly MS is "Absolutely behind OSS".

I work in the public sector in Denmark, and we’ve delt a lot with both IBM and Microsoft over the past 25 years.

Microsoft has been one of our best partners, including for the open source software we run, especially since Azure became their mission.

IBM has been one of the worst, so bad that I’d dread making any deals with them ever again.

I wouldn’t say MS is fully behind OSS though, they contribute a lot these years, but their main goal is still to sell you Azure. I think they won the hearts and minds with .Net core though, I mean VSC is the best ide and typescript is typescript, but the future of a lot of web programming lies within .Net core.

You are likely coloured by being in the one marked where .Net became significant and that’s mostly bacause the only accepted altilernative in Denmark’s monopoly friendly procurement systems were IBM mainframes or Oracle solutions.

Everywhere else RedHat/Jboss won the game and is being replaced by new JWM languages rather then node or .Net though node hides in strange places like the latest SAP framework.

Frontend/native .Net apps are fastly becoming extinct.

MS pretend love for Linux is more an acknowledgement that no one wants dotNET on IIS or anything windows centric in the cloud than any genuine love for Linux so they kind of have to pretend to like Linux workloads and unix tools if they want azure to be more then an niche product.

In many European countries it mostly boils down to Java vs .NET, depending on the business sector.

And by Java I really mean Java, with alternative languages being done by clever consulting companies, which sometimes I get to rewrite back to Java.

Oh we have JBOSS in our stack, it’s what handles our service bus. I’d rather we didn’t though, it’s really hard to find JBOSS developers/maintainers for public sector pay checks.

We used to see a lot more of it from our suppliers, but it seems to be rapidly going extinct. Possible because there just aren’t a lot of JBOSS developers/maintainers in general.

I am coloured by my environment of course, but I do work in software cooperatives with 97 other muniplacities, as well as a few European communities and I don’t see anything to indicate that .Net, JAVA and PHP won’t remain the dominant techs in Europe for the foreseeable future.

I like node.js, I use it for hobby projects and I genuinely think graphql is a lot better than rest APIs (and there isn’t a graphql adaptation for .Net that isn’t bad), but I just don’t see the adoption anywhere outside of what you hear from American startups.

And again, I didn’t say .Net would rule all web development , I said it would be important, and if the European public sector continues to run on .Net then it’ll continue to be a billion dollar industry.

dotNET will stay around just like the mainframes but it’s not a growth market nor the worldwide norm for enterprise web backends.

I work on legacy platforms so I know there is good money in dead technology. But that don’t make it the future.

I just don’t see any legacy codebase being rewritten as dotNET and a similar amount of new greenfield dotNET projects as new Perl project being launched due to NETs heritage as a windows component.

I’m not sure what gave you the impression it wasn’t growing, because it certainly is in Europe.

Come to Europe, plenty of .NET greenfield projects across the continent.

If I were a Perl consultant I’m sure I would say the same about Perl and my region and it’s not that long ago that IBM stopped claiming the future was still the mainframe.

The most available job in my country for any programming language is C#, followed closely by JAVA. On third place is PHP. Fourth is Sharepoint and RPA. Around half the number of the C# jobs include some kind of JS requirements but almost every JS job uses a different backend than node.

There is one fullstack JS job. Three JBOSS jobs and six DJANGO jobs.

This isn’t unique in Europe.

As polyglot developer, .NET is just one of the many tools on my toolbox, just look for yourself on European job boards.

> the future of a lot of web programming lies within .Net core

How is that, really?

How is it not? I mean, we’re a muniplacity, we operate around 500 IT systems, that are all moving toward becoming web apps in some form.

The core tech behind these is in 95% of the cases either .Net or JAVA.

Our in-house development has moved from .Net to JavaScript, mainly because we’re small and if our front end had to be JS then our backend might as well be, but now you have something like Blazor.net emerging, allowing for full stack C#, of course we’re going that route.

I didn’t say all, but I frankly think it’s obvious that .Net will play a big part of web development future, considering how big a part it already plays today and considering how Microsoft is moving it forward in all the right ways.

Historically many different companies and frameworks have targeted the web platform and almost always they lost to "plain old Javascript".

That trend is going to continue, specially considering WebAssembly.

But as of now there's no sign of Javascript becoming less dominant as almost all the innovation is in the land of React, Webpack, Babel, etc.

I mean, what you're claiming "will emerge" is already there in form of Typescript, Node, React, Webpack etc and has a pretty good traction.

Does it JS really have that much traction on the backend? Maybe outside of Denmark, but our little department is actually one of the few places which uses Node for serious backend application in the entire country. At least that I know of.

I mean, I can go on job databases or LinkedIn right now, and there isn’t a single full stack JS job available in my entire country. There is a lot of JS including jobs of course but they all require you to also/mainly do C#, JAVA, PHP or Python because JS is almost exclusive used on the frontend.

Don’t get me wrong, I actually really like the JS environment. There’s a reason we moved to it, but it’s not like it doesn’t have its flaws either.

I think WASM will absolutely change web development, but I think it’s already made it to the is, I mean, we’re launching our first minor Blazor app this week, and it’s something we typically would’ve build with vue and graphql Apollo, but now it’s all C#.

The world of enterprise typically moves slow though. We’ve recently bought an on boarding system that’s made with web forms for instance. You may laugh at that, but the truth is, at least in my part of the world, that JS hasn’t seen that much adoption outside of hobbyists. Eventually these companies are going to upgrade their client sides, but would you pick modern JS or full stack C# if you were coming from web forms? Hell even if they go vue, react or angular chances are they’ll still use .Net on the backend, as that seems to be the trend pretty much everywhere except for us.

Does it JS really have that much traction on the backend?

I cannot speak to the whole industry as I moved to the states a year ago and my view is quite limited. But for sure Node has a lot of traction and usage in here.

However, generally, I feel developers live in their own echo chambers. For example I personally am very much connected to JS people and Linux/Open Source people. I rarely meet any ASP or Java based stacks.

I think it also depends on who you are developing for. Governments and enterprises for sure use more Microsoft or Java based solutions while startups and private companies are more "bleeding edge".

So to answer your question, I think you're right. If an organization is using WebForms, their most probable choice for upgrade is the newest offering by Microsoft. That's where they have already invested it. collectively.

But other stacks have a lot of users too. And they are not going to switch to .NET even if it becomes FullStack.

This Google Trends results [0] are interesting. Not sure if they tell us anything meaningful or not though.

[0] https://trends.google.com/trends/explore?date=all&geo=US&q=....

I manage though, so I get to do the contracting. Mainly our software is enterprise, and it’s always either .ner or java, these days typically with an angular front.

We’re part of several muniplacity driven OSS communities though, where we buy OSS development from small startups and take ownership over project management as well as the codebase.

None of the startups are big on node.js, it’s nostly python, .net, java or php because that’s what they teach at the universities and it’s what they work with in their free time.

I think the node.js environment is great, like I said, but I don’t think it really has much adoption in Europe.

A lot of (older) sites run ASP with some prebuilt or compiled c#/vb to js on top of that. No programmer would have to do JS if they want to.

The technology is flawed in my view, but it does work and has no active js development needed.

Note again: you can, but you don't have to.

Similar experience here in Norway...

If you are including the boot loader, don't leave out the file systems.

Meh. They have a bad reputation due to calling open word source software cancer the new windows business model.

Also, while their engineers are certainly smart, their software seems very crusty (the down side of infinite backwards compatibility) and usually doesn't play well at all with existing open source software. Thus: Mostly useless.

Amusingly, they're damned both ways at this point, since many of the Windows 10 complaints have, in fact, been breaks in backward compatibility.

The early 2000s called and would like their objection back. Microsoft isn't that same company any more and haven't been for a long time. Sure, some of their technology is pretty awful but they definitely changed their tune around OSS.

So we can merge exFAT support into the Kernel now? Microsoft loves Linux, right?

Mostly useless projects though for people that want speed and Linux-native open source.

On the other hand it might mean the end for SystemD!


Debian and Ubuntu still use it.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact