Linux Foundation Launches Valkey As A Redis Fork

[Jaxad0127]

Where does anyone see anything that says this is no longer open source?

Due to the Redis licensing changes, Valkey is forking from Redis 7.2.4 and will maintain a BSD 3-clause license.

BSD is an open source license.

They're referring to Redis, not Valkey.

[ehansin]

I'm disappointed they went to that and then didn't go to the next logical step of Valkeyrie

That will be the fork of Valkey

[jokeyrhyme]

Has un-open-sourcing a project ever worked for anyone?

I'd love to see a study on this

I've often thought that source code isn't the most valuable asset, it's the people who are experienced in enhancing, maintaining, and adopting that source code

So a naive fork without any of that human expertise probably wouldn't last very long and would (hopefully) not see much adoption, although there are some big names involved here so I don't think that's going to be a problem for valkey

And for the window of time where the two projects are compatible, we'll probably see projects use redis for self-hosting and local development/testing whilst simultaneously using valkey when cloud-hosted, so there's some cross-pollination of users there

But I still doubt very much that this move translates into profit for the Redis folks, however much they might deserve it

[jokeyrhyme]

And for the window of time where the two projects are compatible, we'll probably see projects use redis for self-hosting and local development/testing whilst simultaneously using valkey when cloud-hosted, so there's some cross-pollination of users there

This might also harm the adoption of any feature that isn't implemented compatibly in both valkey and redis, making it difficult for either project to entice paying users with new features

[CommunityMember]

I've often thought that source code isn't the most valuable asset, it's the people who are experienced in enhancing, maintaining, and adopting that source code

So a naive fork without any of that human expertise probably wouldn't last very long and would (hopefully) not see much adoption, although there are some big names involved here so I don't think that's going to be a problem for valkey

If you go through the recent commit history of Redis you will likely see that the non-Redis employed contributors appear to be overwhelmingly employed by the big cloud providers and large customers (which is not at all surprising, as large users are likely to see problems due to scale and they want to address those issues).

It will be interesting to see how Redis uses the Valkey commits (by developers that used to generate PR's in Redis itself) going forward.

[timrichardson]

You can provide support/warranty for a fee, and still remain GPL.

The business model of Redis Labs, which was sold to investors, was
+ Use an open source licence for credibility and to rapidly acquire customers, and also to suppress competition from other open source projects
+ Convert a share of these customers into customers who pay for a hosted redis service
+ Convert some customers into those using closed source extra features


So the way redis labs wanted to make money was to compete as a hosting business. People say AWS exploits open source by offering it as a hosted service but it was Redis Labs that chose to compete with AWS for hosting dollars.

the idea of charging for service and advanced features, the open core approach, makes more sense to me. But Redis Labs was already doing this. I guess too many users were happy with the open source functionality.

This means that redis is not valuable enough to earn lots of money. No surprise there, it's just a single threaded key-value database. This type of software is the sweet spot for open source.
There was always going to be an open source something like redis.
it's a shame that the Linux foundation went for BSD but it gives choice I suppose, considering there's redict.
I think I will first try redict.

[fahrenheit]

The issue in my opinion, as someone said already, is that cloud providers (mainly aws, azure and gcp), take these open source (usually community versions) products and offer them as part of their solutions while not giving back much to the vendors. Additionally these providers have their own roadmap which may not align with the vendor.
From the top of my head I remember, MongoDB (of which aws, gcp and azure offer different solutions with varying degrees of api compatibility), Elasticsearch, PostgreSQL, Kafka, etc.

One may think that the cloud providers are offering added value, but in fact they usually offer minimal differences and older versions and you will usually pay twice for the privilege of having a cloud ui to manage some parts of the software while the provider support is only available on the higher tier where you pay 5x more.

Vendors try to change the licence to protect themselves and in most cases earn more money to their shareholders, but that money will also pay for the developers that actually develop and mantain the code. Independent developers are enough for small projects but for large projects you do need paid professionals.

Such is the world where we live in.

[timrichardson]

Has un-open-sourcing a project ever worked for anyone? Especially a project this big with this many stakeholders.

Sure, they'll probably get a little bit more short-term profit, but that will run dry as Redis loses both its installbase and mindshare.

Red hat?

Actually many businesses use open source to make money. They don't sell open source directly, they use it build added value products. They might be selling cars, mobile phones or supercomputers, or added value software .
it's like timber. You don't see it from the outside but most houses are built from it.

Except of course that builders buy the timber. For open source, say you want to launch a phone with some advanced feature, like a folding screen. You have to provide an OS which supports two screens. You could write an entire OS or you could take an existing OS and add the extra bit. That's much cheaper. You just saved millions of dollars. So that's money made by open source. Probably this new phone would not even launch of it required a proprietary OS, and if it did it would have much higher costs to recover. A competitor that used OS + some contributions would have a big advantage, in time and cost.

You have to make your software contributions available to your competitors but you are competing on advanced hardware so that's ok. When lots of businesses all make the same decision you have a sustainable open source project. And now Samsung is locked in to open source. If they want to later make a proprietary os, now they have to reimplement an entire OS AND their contributions they already paid to develop. Every additional contribution makes further contributions more likely. Sustainable.

How
do developers make money? Well Samsung is paying for this coding. That's it. It's not very exciting. An open source company can't be a unicorn because open source means you have to give the software away.

Open source is valuable in enabling other things. It's not going to ever capture much value in its own IP. You can't sell open source software. Investors who fell for the Redis Labs business model (or Docker's) were silly.

[fitzie]

this other redis fork is now probably DOA. They tried to switch from BSD to LGPL https://redict.io/

[dwagner]

Red hat?

IBM will certainly milk some companies for a while while their incompetence keeps them in vendor-lock-in. But every company I have contacts into, both big and small, that has used RedHat in the past, has either already migrated away from it or is in the process of doing so.

Regarding Redis, the effort for companies to switch to the still-open-source fork is very, very low.