Slashdot Mirror
Red Hat Rejects MongoDB's 'Discriminatory' Server Side Public License (zdnet.com)
An anonymous reader quotes ZDNet:
MongoDB is an open-source document NoSQL database with a problem. While very popular, cloud companies, such as Amazon Web Services (AWS), IBM Cloud, Scalegrid, and ObjectRocket has profited from it by offering it as a service while MongoDB Inc. hasn't been able to monetize it to the same degree. MongoDB's answer? Relicense the program under its new Server Side Public License (SSPL).
Open-source powerhouse Red Hat's reaction? Drop MongoDB from Red Hat Enterprise Linux 8. Red Hat's Technical and Community Outreach Program Manager Tom Callaway explained, in a note stating MongoDB is being removed from Fedora Linux, that "It is the belief of Fedora that the SSPL is intentionally crafted to be aggressively discriminatory towards a specific class of users." Debian Linux had already dropped MongoDB from its distribution....
The business point behind MongoDB's license change is to force cloud companies to use one of MongoDB's commercial cloud offerings. This hasn't worked either. AWS just launched DocumentDB, a database, which "is designed to be compatible with your existing MongoDB applications and tools," wrote AWS evangelist Jeff Barr.
Open-source powerhouse Red Hat's reaction? Drop MongoDB from Red Hat Enterprise Linux 8. Red Hat's Technical and Community Outreach Program Manager Tom Callaway explained, in a note stating MongoDB is being removed from Fedora Linux, that "It is the belief of Fedora that the SSPL is intentionally crafted to be aggressively discriminatory towards a specific class of users." Debian Linux had already dropped MongoDB from its distribution....
The business point behind MongoDB's license change is to force cloud companies to use one of MongoDB's commercial cloud offerings. This hasn't worked either. AWS just launched DocumentDB, a database, which "is designed to be compatible with your existing MongoDB applications and tools," wrote AWS evangelist Jeff Barr.
There has been more discussions about monetization of open source projects. While coding as a hobby and helping projects as a philanthropy work is very good, long term stable projects need continuous funding.
This has worked for RedHat and other enterprise oriented companies with their support contract offerings. It also worked nice for existing companies -- including even Microsoft -- which uses open source partially. However if you only have a single offering, like MongoDB the situation was not as clear.
And it is not AWS's fault that they don't want to pay per-seat licensing fees: https://www.techrepublic.com/a...
Who cares? What system administrator cares what their says operating system in terms of databases?
From the article:
No, the business point was to prevent companies like Amazon from hosting MongoDB [theoretically] as-is, like they do with other products like Kafka and Elasticsearch, which have subtle changes and limitations compared to self-hosted instances, but they are largely good enough. In the case of AWS, they were apparently successful enough to force AWS to host a database with a MongoDB-compatible API rather than actually hosting a modified copy of MongoDB's code.
The obvious hope of that situation was that this would prevent those companies (AWS, IBM, Microsoft Azure, etc) from hosting compatible-enough layers, but that has proven unsuccessful with AWS' announcement (as well as CosmosDB on Azure, which was announced awhile before the license change and is closer to AWS DocumentDB than MongoDB).
As an open source developer and fan, I do not understand the hate toward MongoDB's move. It is not exactly an attractive move, but their hands have been forced by the likes of AWS and Azure taking without giving back. They've developed a successful, large business around their database. But because it's open source, Azure and AWS are going to get the lion share of profits without giving back a line of code.
So one vendor can't make an "evil" license with vendor lock-in. I expect systemD and chromium will be the next projects to do something like this. And before you say fork, i say spoon.
So one vendor can't make an "evil" license with vendor lock-in.
If it truly is open source then there is no vendor lock-in. Anyone can fork the code and then develop it and distribute it themselves which I suspect is exactly what Amazon is doing. If you can't do that it is closed source. Open source means more than just being able to see the source code.
Source
Fair and balanced take, overall. Even sympathizes with why Mongo did it, even if not in agreement with the ultimate decision.
AWS is not breaking any rules here, but folks need to seriously look at it from a long term sustainability model and not necessarily go with AWS.
So MongoDB Inc. says they want to have money from companies that host sell the service of of hosting their database for others, and those companies say "thanks, but no thanks, we'll rather use a different database, then".
That's a very simple decision both parties have the right to take, and it's all good now.
I remember when the company I work for had to decide whether to use RHEL or another Linux distribution, and since we deemed RHEL way too expensive for the little added value they offer, we went on to use CentOS (on thousands of machines). That was a similar situation where seller and potential buyer concluded their respective valuations were just too different for a deal.
Mongo/Document DB's so lame..
AWS has been historically ripping off open source projects, forking them, and building new products without any attribution or support. The premise that they are the best cloud provider is kind of tainted to me, since they aren't building anything from scratch but rather, taking the work of thousands of volunteers and incorporating it into a paid for service. Examples would be in addition to DocumentDB (an ironic name since Azure called their NoSQL offering that first), Redshift, Maria, and even their hypervisor.
Then you have Mongo who builds a product and markets it like it's the cure to cancer, and early on had a lot of success in convincing people they needed it (they mostly don't -- read up about Diaspora when you find time). Now that they are in the position of being outed for their proprietary connection interface (which is what makes AWS offerings compatible) they are changing the licensing. I don't think Red Hat's move hurts them as much as people don't want to maintain a fucking Mongo cluster any more.
So who do I feel sorry for? Fuck both of them. Azure and GCP are better cloud providers anyway -- Azure for the enterprise, GCP for 'startups' that don't need the enterprise tooling. If you start your business on AWS it's because you ride the bandwagon either have enough money to burn, or are stupid.
The price is always right if someone else is paying.
Since MongoDB’s functionality seems to be easily replicated/replaced, this seems to suggest there’s too many possible offerings out there for them all to be self-sustainable.
AC comments get piped to
...wasn't Web Scale.
Yep, the sound of a foot being completely blown clean off!
Or you can monetize by entrapping your customers and charging them excessively. This works (at least it does for the seller) if you have a monopoly or near-monopoly (e.g. cable companies). But if you don't, it just makes your customers flee and switch to someone else's product. That's what we seem to be seeing here.
Another thing long term stable projects need is a way to keep tabs on what customers want. What new features were helpful and should take priority? What changes were made that actually annoyed customers rather than helped them?
I haven't seen open source address that, when it's the most important thing you lose when you make a project open source instead of pay software. With pay software, the amount customers are willing to pay for your product signals to you how much they like or dislike what you're doing. It creates a direct positive feedback loop - you add something customers want, more of them buy your product, creating a greater incentive for you to add more things customers want, making more of them want to buy your product, etc. That it also solves the funding problem is just gravy.
Too many debates on Permissive vs. Copyleft hinge on ethics and morality, on what *should* be the outcome. Far fewer look at the practical state of affairs and the forces involved, in other words, what *will* be the outcome.
The simple reality is that today the Copyleft movement is not nearly as strong as it was decades ago, and continues to weaken (not to be confused with the Open Source movement as a whole, which is stronger than ever). Some of the reasons include:
* Companies increasingly embrace (Permissive) open source, but shun away from Copyleft due to not wanting to cascade open-source their own code (again, I'm not saying this is right or wrong, just stating a fact). Decades ago, an enterprise wrote everything in-house, so it was a single company vs. the entire open source community. Now, companies aggressively leverage (Permissive) open source, so the cards aren't nearly as stacked.
* Some of these companies (even if not all) contribute back to the tools they use, even if purely for selfish reasons such as leveraging community maintenance, preventing forking, or ensuring that "their" extension makes it way as a "standard". Historically, companies used to shun away from standards in favor of a proprietary implementation, but now they begin to realize that it's far more profitable to *make* your own design and implementation the standard - even if it means open-sourcing it (Permissively) in the process.
* Developers, when working on open source software, tend to want to work on things which will get them the most experience and recognition when they find commercial work. Working on related software that is used by corporations makes it easier to learn, compared to tools which corporations intentionally avoid using. Activism is a privilege for those that can afford it. Most people just do what they need to get by, and developers are not an exception.
* Many developers just don't care about this debate at all, and will use whatever's more convenient. Not having to worry about licensing at all is easier than having to worry about it (e.g less cognitive/administrative overhead to make sure everything is properly followed), even if you're not against Copyleft in principle.
All this creates selective pressure in favor of Permissive licenses (corporations, corporate money and influence, developers that want to work on things used in the corporate world). In the other corner is Copyleft which has *no* force behind it other than activism and personal beliefs of some (but by no means all, or even most) open source developers, and the (increasingly weakening) Copyleft viral effect. The Copyleft movement hoped for an "avalanche effect", but we're now seeing a "reverse-avalanche" where the corporations and developers increasingly engage in a Permissive license development loop, whereas Copyleft software is increasingly marginalized.
The entire premise of Copyleft is to get open source developers to stand as one, leverage strength in numbers, and create a reality where no viable alternative for a corporation exists than to use Copyleft software (and open-source its own software in the process, creating an avalanche effect) than to build everything in-house. This reality is *shattered*. Developers are fragmented even between themselves due to the above-mentioned reasons, and when the corporate world, on top of that, strongly supports one side over the other, the "strength in numbers" premise falls apart. And without this premise, Copyleft does not have sufficient leverage to force companies to use it. For almost any tool that exists as Copyleft (and that was developed in the last decade or so), there's a comparable (if not better) tool with a Permissive license. *Especially* so in the Webdev world, around which this whole SaaS debate is now happening.
Consider LLDB (largely superceding GCC). Postgres (largely superceding MySQL/MariaDB in the web world). Most web languages (JS V8, Golang, Rust, Scala, Python, Ruby, Elixir, PHP, .NET Core). All the Apache software. Nginx
Up until 5 years ago I could make an open source project and make money off it by providing services. Engineering liked this because they felt safe that they weren't locked into a single vendor who might go bankrupt. Accounting liked the word free even if the service costs eventually cost more. The model seemed good for everyone.
Then came AWS. The software users are willing to pay for something they feel is tangible, computing time, storage and support. AWS is amazing for all three but then AWS became the support for the open source projects. The end customer started paying AWS for support and not the companies developing the open source project. Also AWS support is far better than what any single open source company can offer. The open source financing model for things like MongoDB went from workable to impossible in just a few years.
Others have posted plenty of long-term stable open source projects with no direct funding to the project. So what's different about MongoDB? The MongoDB company, which is trying to make money from the project. *Making money* from open source requires that someone pay money, for something.
Plenty of open source projects work fine without anyone funding the project in any significant way - a developer who wants or needs a feature codes it, and makes a pull request. That developer might be at work or at home, but nobody pays the project anything.
The question is how can the Mongo company get money from someone. There are many options. An open source project called Moodle has an interesting method. Their project page lists companies who provide Moodle services - hosted copies, custom development, training, etc. "Official Moodle partners", I think they are called. Those companies are allowed to use the Moodle trademark in their marketing if they want to, and they pay a percentage of revenue for being listed as an official partner.
Platinum level -> over 100,000 instances / year -> $1M donation
Gold level -> over 10,000 instances / year -> $100k donation
Silver level -> over 1,000 instances / year -> $10k donation
Supporter -> less than 1,000 instances -> at your discretion
How do you find out whose running this many instances?
Phone home for update checks (enabled by default) and explicitly state that statistics are tallied to support and fund the open source software.
What happens when BigCloudVendor is caught running xx,xxx instances but doesn't throw you a donation? Call them out.
Mongo just pawn in game of life.
Why not?
Sent as ripples into the electromagnetic field. No single photon has been harmed in the process.
MongoDB is Webscale! MongoDB no more Websell!
Literally the AGPL ensures all source related to a web service is provided to any customer free of charge. It's the perfect rent seeking license for open source projects on the web, and has existed in various forms for 20 years.
If MongoDB was under it, they would be free software compliant AND if Amazon violates their code in any way, they have them over the fire. It won't solve their current situation, but it will make Amazon have to step carefully in the future.
Anyone working on web oriented open source projects needs to consider the same. If that doesn't work for you because your project is a support library, talk to Affero and FSF about forming a new ALGPL for that use case.
Then license for LGPL or commercial purposes.
EInit, by Emacs.
Up yours!
Doesn't matter what ethical system you want to blather on about, none have any bearing on what business does. business knows no morals nor ethics. its just business, nothing personal. Make 'merika Gree^H^H^H^HInnovative Again.
Who IS Max Webster?
> Debian Linux had already dropped MongoDB from its distribution....
That certainly doesn't appear to be true, at least at the time of this writing.
Buy this hipster a salted caramel latte!