Redis Inc seeks control over future of Rust redis-rs client library
devclass.com
devclass.com
Salvatore Sanfilippo (antirez) started Redis and developed it by himself from 2009 to 2015, gaining massive popularity and building a large community in the process. It was FOSS the entire time.
A separate VC-backed company called Garantia Data used to make money by offering a hosted version of Redis. That company changed its name to Redis Labs in 2014 (and eventually just Redis), likely themselves violating antirez's Redis trademark at the time.
They then hired antirez in 2015 and started officially sponsoring the project.
From there began a slow transformation of Redis from a community run FOSS project to a proprietary locked down service. The company also managed to acquire full rights of the Redis trademark and project stewardship from antirez after hiring him and then finally kicked him out in 2020.
So yes, antirez started it. He owned the trademark and gave it off to redis inc. and was compensated for it. I am not seeing why this has to be controversial.
I don't like what redis is doing. But they're within legal rights.
But this is a takeover that is slowly draining the value from the community and directing it to private pockets. E.g. Redis is now source-available.
There are still compatible alternatives: https://valkey.io/ (C, a direct Redis fork) or https://keydb.dev/ (C++, an evolved Redis fork), both BSD-licensed.
I wish RethinkDB was more alive :-\
I think the best option against this action by Redis Inc is to move to valkey and use it to incorporate features that Redis Inc considers "enterprise licensed features". If there are big players in valkey, there should be a move to setup a foundation for it - so that it doesn't get taken in to be source available again.
In theory Redis could have had a foundation. It didn't. Instead the project owner opted to sell it to a corporation. This doesn't happen by accident.
What indeed has a foundation is Valkey[1]. It's backed by the likes of AWS, Google Cloud, Digital Ocean, etc.
[1] https://www.linuxfoundation.org/press/linux-foundation-launc...
I think you're fooling yourself into believing that Redis was ever a community-owned project.
The truth of the matter is that the guy who developed Redis ended up selling Redis to a corporation. Since then he bailed out, and the corporation that owns Redis is now going through great lengths to monetize it.
The faster you forget about Redis and switch to alternatives, the better you'd be.
How did that happen? He must have given/sold it to them, right? I remember him making an announcement that he was done with Redis and stepping away from involvement.
Created a PR to add this into the context.
Feel free to comment on it.
Also, calling other bystanders to add other missing pieces to the history.
My rule of thumb now is that I now consider any project that has a pricing page OR requires copyright assignment/CLA to a for-profit company to be effectively proprietary and just using open source as a marketing technique. That doesn't mean I won't touch it, but like with proprietary software, I'll evaluate it against the risk that the price will probably be jacked up in the future.
QT has contracts with KDE around the open source version which gives KDE peace of mind. I use QT in a commercial product - we have some useful contract terms with QT that are not public and I can't talk about them.
> Thanks everybody for the feedback. Speaking on behalf of Redis Inc., we want to find a way to collaborate to best support the community and our customers. The objective is to ensure predictable releases for a Rust client library, manage issues and escalations promptly, as well as support the best we have to offer without forking the library and competing with the client library project. After discussing this with @nihohit in this thread and based on the whole conversation, we want to work together. We have already identified initial areas from which we could start.
> We have no issues keeping the project name as it is without a transition to Redis. We also have no problems with continuing to call this library "redis-rs". There is no intention to claim ownership of the client library's name, source code, or the crate’s package registry.
Rust seems to just attract drama sometimes, the other client library owners dealt with the company without blowing up?
> the other client library owners dealt with the company without blowing up
A lot of the other client libraries are already under the control of Redis Inc. The Python client, one of the popular Java clients, the Go client and the nodejs package all live in the Redis Inc Github organization.
As the author of that issue I'm assuming if there was drama, then it was up to me. However I did not intend on causing one, but to discuss this issue with active maintainers of the crate as well as to understand to which degree valkey support is needed by users for the crate.
That this has created a discourse that goes beyond that was not intended.
I think the thing started off fine and reasonable, but if you go down the comments it takes a turn towards cynical and antagonistic where people are assuming the worst. Which is basically the point of my comment, rust related things seems to have these weird blow ups.
Some quotes
> Redis team has the required Rust proficiency, nor that they actually care about maintaining this crate
> Concepts of a plan eh?
> Of course they don't have a list of missing features, it's not about features. It's about taking control of a ecosystem that's collasping under them because of widly percieved-as shady license rug pulling.
In bold too
> What you care about is your customers, not the community or any contributors.
Then there's headlines like
> Redis Inc seeks control over Rust Redis-rs library, talk of trademark concerns
Its overall inflammatory, when the intention from the emails shown seem fine and the goals seem clear.
Thank you antirez, mitsuhiko, and mortensi for working to resolve this amicably!
1/10 are using it as a hope-for-the-best "queue" instead of rabbitmq, which is bullet-proof.
The last 1/10 actually use it as a novel "database" but every one of those instances also has mysql or postgres, rendering it completely redundant.
Redis itself was, for a while, a massive open security hole when the above people would put it on the open internet, where it would to quite useful to hackers as a free lua program runner.
Or would you still prefer to build on top of memcached?
If you are running slow queries and caching them into into a k:v store you can just do that with a TEMPORARY TABLE - sql databases are absurdly fast k:v stores.
There are valid use cases for redis in its novel types and things but my comment was trying to express that seeing that usage, in my real life experience, has been rare.
https://en.wikipedia.org/wiki/Embrace,_extend,_and_extinguis...
Redis is risking its reputation in order to solidify its revenue stream in the face is rising threats like Valkey, etc.
If you want something hosted/managed, there’s Upstash Redis (though I reckon they’ll soon have to change the name of that offering).
That's rich considering how they've been actively destroying their reputation by themselves.
The community took steps launching several Redis alternatives, including Valkey - the next step would be also to get rid of trademark in the connectors
As there are number of vendors offering Redis compatible databases those days I think the best approach would be to come up with vendor neutral name for Redis protocol and when Redis, Valkey, DragonflyDB etc could be listed as supported products.
Redis is trying to take over the all of the OSS Redis libraries
We don't need this noise. The code is already written and published. Consider the 'brand recognition' of such exciting tooling as:
* fzf * tmux * ripgrep * exiftool * fdupes * etc.
I am trying to remember why their software became considered ubiquitous for caching and sessions, and I reckon many a framework is busy rectifying this choice, as we speak.
I'm willing to bet most people running Valkey, Dragonfly etc. on the server are still using the Redis clients.
Is this in bad form? What does the guy have to do to convince you that he has to rename the library? It's tough cookies, but if he renames it, and the Redis Ltd. people fork the library and put the fork on crates.io under the redis name, that's what happens. The way it works just isn't whoever gets the name on crates.io first, irrespective of copyright.
I'd think that if the situation were reversed - Random Guy On GitHub Complains About Distasteful Actor Taking Over His Trademark - you'd root for the guy no?
That might not be legally relevant but it is certainly ethically relevant.
I don't know if it's ethically relevant. I'm sure there was someone named Matt Damon before the actor Matt Damon, and maybe that guy was even an actor, but I wouldn't say today's Matt Damon is ethically violating ancient history's Matt Damon.
What is the right rule for abandonware? It can't be, whoever got there first. Anyway. I don't buy your timeline. Redis-rs comes after the name Redis, certainly, which these guys now own. It doesn't matter when these two events you picked out of a hat occurred.