Comment by jamespo
4 months ago
With 100PB clusters being built and not a cent going to them, you can see why minio has gone this route. I wonder if they will be "valkeyed"? Not by AWS presumably.
4 months ago
With 100PB clusters being built and not a cent going to them, you can see why minio has gone this route. I wonder if they will be "valkeyed"? Not by AWS presumably.
That's the open source model. It's entirely predictable that if you provide software at no cost that is capable of running 100PB clusters, that some people will and you won't get paid, because those are the terms that you set.
It's fine to change your mind, but doing it in this way doesn't build goodwill. It would be better if they made an announcement that they would stop creating/distributing images on some future date; I'm sure that would also be poorly received, but it would show organizational capacity for continuity.
If I'm considering paying them for support, especially at the prices quoted elsewhere in the thread, I need to know they won't drop support for my wacky system on a whim. (If my system wasn't wacky, I probably wouldn't need paid support)
There are a few challenges with open-source projects that want to also be commercial entities.
One is obviously knowing what you can add-on that people will pay for; support, for one, but people want more features too. What could minio have built on top of their product to sell to people? Presumably some kind of S3-style tiered storage system, replication, a good UI, whatever else, I'm not sure.
The second is getting people to actually know that that's an issue. I work for Tigera which publishes the Calico CNI for Kubernetes, and one of the biggest issues we have is that people set up Calico on their clusters, configure it, and then just never think about it again. A testament to the quality of the product, I'm sure, but it makes it difficult to get people to even know we have a commercial offering, let alone what it is and does and why it might be beneficial.
I could see the same thing for Minio; even if they have a great OSS product, a great commercial offering on top of that, and great support, getting people to even be aware of it in the first place is going to be a huge challenge and getting people to pay for it is even harder.
It's sad that they went the completely wrong direction and started taking things away from the community to force people to the commercial side of things whether they're willing to pay or not.
I reckon they gave away too much, and are clumsily rowing it back.
Gitlab seemed to do a good job of navigating a community edition as an on-ramp for sales. But it's obviously a lot of work to maintain that edition, and VC must be feeling less geenerous than 10-15 years ago.
e.g. maybe if it were my project I'd have kept back the S3-compatible ACL support and put in something super-basic. Or even cluster support. Right now it feels like they're cutting off everything they can while still being able to call it "open source".
That's a strange mindset, IMO. I'd be pissed if I had to pay $0.10 every time I turned a rachet, and it's weird to expect companies to have usage-based monetization on the tools they've made for others.
An analogy to making a physical tool doesn’t really work because we have to basically describe what software is in terms of exceptions to the analogy.
If I had a ratchet that, every time I turned it, I had to pay $.1, but I’d gotten it for free, but it was basically free to replicate, but the person who designed it did have to spend some significant work on R&D for the thing… I have no idea how I’d price that or how I’d feel.
oh you really butchered that metaphor.
The ratchet isn't what's getting paid in the metaphor, it's the person turning it.
There's always a time-sink cost to a public project.
Anyway, there's definitely a public good argument to turn certain software projects into utilities.
3 replies →
did you buy the ratchet?
that's why you'd be pissed.
If you were given the ratchet and then someone wanted to charge you every time you use it you would also be pissed.
9 replies →
In this example the ratchet manufacturer would be giving them away for free though, and then get pissed when no one volunteers to pay.
Let me introduce you to Splunk and enterprise software in general
You effectively do pay per turn of the ratchet. It doesn't last forever, will eventually break, and so you can amortize the cost of the device over the number of turns you expect it to make to get the per-turn cost.
Software on the other hand does not naturally wear out, in the same way physical objects do.
> I wonder if they will be "valkeyed"? Not by AWS presumably
Almost certainly not, due to the AGPL license. I know Nutanix got into hot water about distributing Minio so I don't think any big shop will fork it.
Nuantrix distributed a version that was still Apache licensed and merely failed to disclose they had made changes.
This is after MinIO asserted that Weka had also stolen their AGPL-licensed code, showing that they extracted binaries from the distribution. They forgot that that 3-month old (unmodified) version was still Apache licensed though.
MinIO generally don't seem to consult lawyers often. They haven't even set up copyright assignment / CLA immediately after switching the license, so technically they are also incapable of selling AGPL license exceptions just like everyone else.
I've done my best to keep MinIO away from most infra I manage, not because of legal concerns but because it was kind of obvious they'd eventually go full scorched earth and either drop images or the source code distribution all together. Maybe now we can all move on to a fork, or SeaweedFS, or Ceph, or literally anything else.
They don't consult lawyers. The CEO husband and wife team get really angry and fire off threatening letters, but I've never seen them consult a lawyer before sending a letter like that or accusing a company of violating a license publicly.
> showing that they extracted binaries from the distribution
Funnily enough, such action is outside of their paid product's EULA.
It’s the sort of behaviour that makes them relying on them even as a paying customer extremely risky.
That just means the fork would also need to be AGPL licensed, and the owner of the fork wouldn't be able to also sell a proprietary version with additional "enterprise" features. And IMO that would be a good thing.
I think it is unlikely a single entity would do that. But a coalition of current MinIO users might get together to create such a project, perhaps under the Auspices of a foundation such as the Linux Foundation. Although, I think that scenario would be more similar to OpenTofu than Valkey.
I am definitely not a lawyer, but as a thought experiment, would Amazon be able to take the AGPL Minio source code, turn it into a managed service, and resell that to customers?
Was under the impression that the answer is yes, they could - with the caveat that they'd have to release the modified source code of whatever backend services are also tied into the Minio source code. For example the AWS control plane that would launch customer instances of Minio, monitor it, etc would also need to be open sourced?
If they charged a cent, would people adopt it in the first place?
They still got paid for those free users. Via investments. Cash is cash. I don’t KNOW what the RIGHT business model is, I don’t run MinIO, and neither do you.
maybe they got paid in exposure
Wait until you find out how much compute is being run on Linux without a cent going to Linus.