I have yet to hear a single positive story about this Redis Inc... it's like a giant company full of only assholes. Story after story is just "wow, these people all suck"
Redis the project was essentially taken over by a company that had nothing to do with its development.
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.
They paid antirez and I'm sure compensated him for his efforts on redis. I haven't heard antirez being "kicked out". There may have been a separation of ways when redis inc. decided to not be truly open source, but I haven't heard of them being abusive or unethical with antirez.
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.
Yes, nobody says that there's something illegal here. Were it so, Redis is high enough profile project for someone to take a legal action.
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.
My point is the lesson here should be to wary of tools that don't have a foundation backing them. I have nothing but the highest respect for antirez. But redis never had a foundation. That makes it susceptible to be source-available. The same with Elastic search.
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.
> 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.
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.
No one is saying what they did was illegal, but you'd have to do a lot of mental gymnastics to make a case for it being ethical and in the spirit of open source.
> The company also managed to acquire full rights of the Redis trademark and project stewardship from antirez after hiring him
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.
Impossible to know as an outsider. They could have tricked him with false promises ("we'll take good care of the project and always put the community first, trust us"). Or he could have decided that the check was big enough and not really cared beyond that.
That seems inappropriate to me for a README — at least in its current, somewhat inflammatory, form. Is the commit itself an act of protest or is this a genuine attempt at getting it pulled?
It's not right to say that Redis "had nothing to do with its development."
They were a major contributors and were managing the project for many years. The problem like many other open source project is that the innovation is shared with competitors who happen to be huge corporates. What would you do if for every line of code you add your competitors get X10 customers than you?
PE/VC-backed bait-and-switch takeovers of "open source" projects have cost me a significant amount of time and money over the past few years.
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.
The important part is the contract the company signs for you. Contacts generally are enforceable in court and lawyers know of standard provisions for weird situations (what if the company goes bankrupt)
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.
If we take the maintainer by his word (and I don't see why we shouldn't) then this was very necessary drama that caused Redis Inc to back off.
> 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.
> Comments are all reasonable and there was no reason for the drama in the first place...
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.
> 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.
Unfortunately ever since the relicensing the situation in the Redis community is loaded. I have seen discussions in other repositories around Valkey and the discourse is not much different.
8/10 uses of redis I have encountered in the world were people using it as a slower memcached.
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.
Isn't it good for building and querying custom indexes? Doing that with mysql or postgres is very difficult or impossible depending on hosting solution.
Or would you still prefer to build on top of memcached?
VIEWS and MATERIALIZED VIEWS probably are a better version of what you are talking about?
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.
My theory is Redis is trying to take control over all popular libraries that interface with it so it can break protocol level compatibility to force vendor lock-in
That would push everyone to valkey. They want to add proprietary features supported only by their server and client. That's the extend part of "embrace, extend, extinguish".
Lets face it Redis Inc owns Redis trademark, same as it owned Redis software copyright and we should expect it will enforce it.
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.
The real value of open source code is that it should be able to be fully decoupled from trademarks. Much like OpenTofu, we shouldn't be caring too much about what private entities are trying to do to disrupt the community. Fork the code, change the names, and move forward together.
We don't need this noise. The code is already written and published. Consider the 'brand recognition' of such exciting tooling as:
I have no need for Redis in my life. There is nothing unique it provides in 2024, and they have no special sauce I would consider getting hooked-on (locked into).
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.
Redis Ltd. probably parent of Redis Inc. owns the trademark. It isn't complicated, they can go around and ask people to change the names of their stuff away from Redis.
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?
redis-rs has been around since 2013, so before redis inc. was called anything related to redis (the were called Garantia data) or they hired the redis creator (in 2015) or bought the redis trademark (in 2018).
That might not be legally relevant but it is certainly ethically relevant.
> 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.
I have yet to hear a single positive story about this Redis Inc... it's like a giant company full of only assholes. Story after story is just "wow, these people all suck"
Redis the project was essentially taken over by a company that had nothing to do with its development.
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.
They paid antirez and I'm sure compensated him for his efforts on redis. I haven't heard antirez being "kicked out". There may have been a separation of ways when redis inc. decided to not be truly open source, but I haven't heard of them being abusive or unethical with antirez.
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.
Yes, nobody says that there's something illegal here. Were it so, Redis is high enough profile project for someone to take a legal action.
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 :-\
My point is the lesson here should be to wary of tools that don't have a foundation backing them. I have nothing but the highest respect for antirez. But redis never had a foundation. That makes it susceptible to be source-available. The same with Elastic search.
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.
> But redis never had a foundation.
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...
> 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.
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.
No one is saying what they did was illegal, but you'd have to do a lot of mental gymnastics to make a case for it being ethical and in the spirit of open source.
> The company also managed to acquire full rights of the Redis trademark and project stewardship from antirez after hiring him
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.
The only public record I've seen him talking about it was in https://github.com/valkey-io/valkey/issues/544, where he mentioned having sold the copyright.
Impossible to know as an outsider. They could have tricked him with false promises ("we'll take good care of the project and always put the community first, trust us"). Or he could have decided that the check was big enough and not really cared beyond that.
https://github.com/redis/redis/pull/13670
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.
That seems inappropriate to me for a README — at least in its current, somewhat inflammatory, form. Is the commit itself an act of protest or is this a genuine attempt at getting it pulled?
excellent summary (not kidding); thank you for it because i hadn't understood it until what you wrote
It's not right to say that Redis "had nothing to do with its development." They were a major contributors and were managing the project for many years. The problem like many other open source project is that the innovation is shared with competitors who happen to be huge corporates. What would you do if for every line of code you add your competitors get X10 customers than you?
PE/VC-backed bait-and-switch takeovers of "open source" projects have cost me a significant amount of time and money over the past few years.
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.
The important part is the contract the company signs for you. Contacts generally are enforceable in court and lawyers know of standard provisions for weird situations (what if the company goes bankrupt)
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.
Shame because I remember the original author being quite well regarded on places like here and Reddit.
He still is
Did he start Redis Inc or got hired to work there? I remember something like that for a popular open source project.
They have some of the same people from the elastic license debacle so this makes sense.
I agree. At the same time, they're just 'doing their job' working at a for-profit company controlling the brand of open-source (core?) software.
Yes, you can simultaneously condemn individual behavior and the system that incentivizes it. Both are bad in this case.
My point is that the blame is on the system moreso than its expected outcome.
Seems like this was resolved with Redis Inc backing off prior to the HN post. From @mortensi roughly 4 hours prior to the post: https://github.com/redis-rs/redis-rs/issues/1419#issuecommen...
> 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.
Comments are all reasonable and there was no reason for the drama in the first place...
Rust seems to just attract drama sometimes, the other client library owners dealt with the company without blowing up?
If we take the maintainer by his word (and I don't see why we shouldn't) then this was very necessary drama that caused Redis Inc to back off.
> 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.
> Comments are all reasonable and there was no reason for the drama in the first place...
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.
> 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.
Unfortunately ever since the relicensing the situation in the Redis community is loaded. I have seen discussions in other repositories around Valkey and the discourse is not much different.
https://github.com/redis-rs/redis-rs/issues/1419#issuecommen... - looks to be mostly resolved at this point, with Redis Inc. simply going to step up its contributions to the open-source version without taking control.
Thank you antirez, mitsuhiko, and mortensi for working to resolve this amicably!
8/10 uses of redis I have encountered in the world were people using it as a slower memcached.
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.
Isn't it good for building and querying custom indexes? Doing that with mysql or postgres is very difficult or impossible depending on hosting solution.
Or would you still prefer to build on top of memcached?
VIEWS and MATERIALIZED VIEWS probably are a better version of what you are talking about?
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.
My theory is Redis is trying to take control over all popular libraries that interface with it so it can break protocol level compatibility to force vendor lock-in
That would push everyone to valkey. They want to add proprietary features supported only by their server and client. That's the extend part of "embrace, extend, extinguish".
https://en.wikipedia.org/wiki/Embrace,_extend,_and_extinguis...
Some cash cows would remain stuck and they are ultimately the ones that would be milked for profit even if 95% of the community leaves
All of this drama is already doing that.
My theory as well. I would almost bet on it.
Redis is risking its reputation in order to solidify its revenue stream in the face is rising threats like Valkey, etc.
Well it's either solidify revenue stream or likely go out of business. And what's a reputation if there's no business to attach it to?
Tangential, I realized I can be satisfied by Redis from 2013. Who are relying on recent Redis versions?
Are there many redis drop-ins alternatives?
Valkey, dragonfly, kvrocks are all protocol-compatible and mostly drop-in replacements for upstream Redis.
If you want something hosted/managed, there’s Upstash Redis (though I reckon they’ll soon have to change the name of that offering).
Microsoft has also been working on one. https://microsoft.github.io/garnet/
Kvrocks is pretty substantially different, from my understanding. It only shares the protocol.
https://github.com/dragonflydb/dragonfly Is a multi-threaded drop in replacment
Valkey is the fork/drop-in replacement from the Linux Foundation.
> "companies do consider protecting their trademarks where their reputation is challenged"
That's rich considering how they've been actively destroying their reputation by themselves.
Previous discussion: https://news.ycombinator.com/item?id=42239607
Lets face it Redis Inc owns Redis trademark, same as it owned Redis software copyright and we should expect it will enforce it.
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.
Alright gang, from now on, all community projects are named Sider.
Well done, for those that missed it: "Sider" is "Redis" back-wards.
Related:
Redis is trying to take over the all of the OSS Redis libraries
https://news.ycombinator.com/item?id=42239607
The real value of open source code is that it should be able to be fully decoupled from trademarks. Much like OpenTofu, we shouldn't be caring too much about what private entities are trying to do to disrupt the community. Fork the code, change the names, and move forward together.
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 have no need for Redis in my life. There is nothing unique it provides in 2024, and they have no special sauce I would consider getting hooked-on (locked into).
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.
Because its very stable, very fast and very well documented / supported.
Valkey
Is there a mature valkey client for Rust?
I'm willing to bet most people running Valkey, Dragonfly etc. on the server are still using the Redis clients.
Just fork the Rust client for Redis before any breaking changes were made.
I don't think there are any at the protocol level. At least not yet.
Redis Ltd. probably parent of Redis Inc. owns the trademark. It isn't complicated, they can go around and ask people to change the names of their stuff away from Redis.
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?
redis-rs has been around since 2013, so before redis inc. was called anything related to redis (the were called Garantia data) or they hired the redis creator (in 2015) or bought the redis trademark (in 2018).
That might not be legally relevant but it is certainly ethically relevant.
> 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.
Analogies are not arguments. Just made up scenario.
Hollywood has very strict rules regarding names. If your name is identical to the name of another actor, you must use a variation or alias.