Can we just go back to libmemcached?
What’s stopping you?
Oh, I have. We use a mix of libmemcached and local file storage - it works great.
The big difference is Memcached is multi-threaded, while Redis is single threaded.
That makes Redis more efficient - it doesn’t have to waste time with locks and assuming the server isn’t overloaded any individual operation should be faster in Redis. Potentially a lot faster.
But obviously by sharing the load across threads as Memcached does, and that cant theoretically allow higher throughput under high load… if your task is suited to multithreading and doesn’t involve a shedload of contested locks.
Which one is a better choice will depend on your task. I’d argue don’t limit yourself to either one, consider both, pick the one that aligns best with your needs.
I told a bunch of people this would happen given OpenSearch. Color me unsurprised.
Obvs it’s about control and money, but it highlights (perhaps badly, but also validly) a need for a mechanism to get funding from profiting entities for open source development, and unfortunately GPL and charity is not it. A well-thought-out license with an expectation of sharing profit from OSS would be a boon for OSS and its (independent) developers. Sure would be nice if FOSS developers got to eat for their trouble (rather than going after yachts). Could be something for the EFF or similar minded, legally aware types to have a chew on, or maybe there are pre-existing works that are not as well known as they should be.
Is it really cheaper for the mega-corps to have their own fork instead of pay Redis for a license? Or are they just capitalizing from loud open-source advocates who are OK working for free for mega-corps because they are principled?
Probably, AWS and Google probably have millions of existing customers using Redis. And AWS and Google are not going to be paying for it themselves of course, but just pass the costs on to their customers.
So they can stick to the old official Redis version for a while, before the license change happened, but at some point someone might find a vulnerability, and patch it in the official Redis, and then everyone that’s stuck on the old version is fucked - it’s a bit of a ticking time-bomb to be stuck on an old version.
So then AWS and Google customers can decide
- “I want to use the latest version of official Redis, and pay x per month per Redis cache” (if the new license allows that)
- Or “AWS doesn’t support a free Redis anymore, but competitor does, so I’m just gonna migrate my infra to a different cloud”
So if they already switch to an open-license fork they can preemptively mitigate most of those risks
That is a different perspective I hadn’t considered. Thank you 🙂
The megacorps will have to pay for the development of their new fork now - or not, if they can find suckers to do the work for them for free, but I doubt that’ll happen. How much that’ll be and how much the customers using the new fork will bring in will probably determine the health and existence of the fork. Unless of course the corps use “fuck you” money to kill redis.
Unless of course the corps use “fuck you” money to kill redis.
I doubt the corps are angry at redis. They just don’t like the new terms.
Nothing to do with anger. It’s just business. They were allowed to leech, now they can’t and the naive, purist opensource community (notice the qualifier purist aka not the entire opensource community) will happily join the ranks of businesses that couldn’t be bothered to donate to Redis despite their money chests. Redis is now a competitor and capitalists don’t like competition. They like monopolies ergo valkey has to become the new #1 and redis has to die.
They were allowed to leech, […]
Services like AWS and Google Cloud offer 1000s of “free software” (like Redis) as a service - like AWS Elasticache - and if you look at the pricing, cache.t2.micro for example, is $0.017/hour, while just a plain t2.micro vm is $0.0116/hour. So effectively AWS is only “leeching” $0.0054 an hour on the Managed Redis that they’re offering.
An AWS managed Redis is just easier, otherwise I’d have to boot my own t2.micro, and install Redis there. I’d still be using official Redis on AWS, because self-hosting is still fine in the new license, it’s just more work for me, because the license doesn’t allow AWS to do it for me anymore.
and the naive, purist opensource community ([…]) will happily join the ranks of businesses that couldn’t be bothered to donate to Redis
It’s funny how people are now siding with Redis. When other companies did something similar (like identityserver4 was FooS, and then they created their new commercial company - and everyone was like “fuck you, you people are sellouts.”.
Most of the time when a FooS project goes commercial, people make a free fork and the commercial project slowly dies
So effectively AWS is only “leeching” $0.0054 an hour on the Managed Redis that they’re offering.
24*365*0.0054 = 47,304$/ year / instance. How many instances are hosted by AWS? How many are hosted by Google? How many by other Redis aaS?It’s funny how people are now siding with Redis
I live in the real world, not some make believe where we live from the gratification of free work alone. It was fine when ElasticSearch did it and it’s fine when Redis is doing it. If you take an opensource project, build a business off of it, make millions, don’t contribute back in a significant manner, and compete with that opensource project’s own product hampering its development, then this kind of license is in no way a surprise.
243650.0054 = 47,304$/ year / instance. How many instances are hosted by AWS? How many are hosted by Google? How many by other Redis aaS?
Formatting it like
47,304$/ yearseems like you’re saying it’s $47k, but it’s just $47. How much would it cost any company to self-maintain their Redis instance?Don’t contribute back in a significant manner
They have multiple people full time employed that are contributing, and how are they “hampering its development”?
If you look at the top contributors: https://github.com/redis/redis/graphs/contributors
- #4 Binbin works for Tencent Cloud
- #6 Zhao Zhao works for Alibaba
- #7 Madelyn Olson works for AWS
- #9 Wen Hui works for Huawei
Soo actually these cloud providers are some of biggest contributes to the project. They’re not just taking Redis and aren’t contributing. The opposite actually lol.
Besides, you’re acting like Redis is some poor little startup, but they’re a company with 991 people (by their linkedin stats). Its like if Oracle would change the MySQL license, and then you side with Oracle “Poor little Oracle, everyone uses MySQL, but no one contributes” - yea no
It’s worth remembering a lot of these megacorps do employ people directly to work on FOSS projects. Here’s a quick and lazy example involving AWS
https://redis.com/blog/redis-core-team-update/ but Red Hat and others do the same.I’m not a fan, and it feels almost as if by employing and embedding people in these projects they look to exert control over them. Realistically, I don’t see that as any different than if they were paying money directly for the same control. Except this way FOSS still has benefits after the license change.
I’d say paying money is not as effective at influencing a project as embedding developers is.
In terms of bang for the buck, I’d absolutely agree. It’s only when a company fully depends on the income of a single client, or closely aligned few, that this becomes a question.
The “mega-corps” have enough money that there’s zero chance this is a money issue. It will be about control.
They don’t want Redis to dictate terms over the software running on their servers. No amount of money will change their mind on that - they wouldn’t accept the new terms if redis offered to pay them billions of dollars.
I hope mega-corps burn billions in the process and redis stays more popular than whatever they develop. They always had the money, but they never wanted to give it to the creators and would rather leech off of them. It’s despicable business and had redis gotten that kind of money before, they maybe never would’ve changed the license. But who knows, maybe they’d have become greedy, just like those fat suits in GAFAM+.
This is the best summary I could come up with:
The Linux Foundation last week announced that it will host Valkey, a fork of the Redis in-memory data store.
This fork originated at AWS, where longtime Redis maintainer Madelyn Olson initially started the project in her own GitHub account.
Olson told me that when the news broke, a lot of the current Redis maintainers quickly decided that it was time to move on.
“When the news broke, everyone was just like, ‘Well, we’re not going to go contribute to this new license,’ and so as soon as I talked to everyone, ‘Hey, I have this fork — we’re trying to keep the old group together,'” she said, “pretty much everyone was like, ‘yeah, I’m immediately on board.”
Redis’s announcement came right in the middle of the European version of the Cloud Native Computing Foundation’s KubeCon conference, which was held in Paris this year.
One area Redis (the company) is investing in is moving beyond in-memory to also using flash storage, with RAM as a large, high-performance cache.
The original article contains 1,171 words, the summary contains 167 words. Saved 86%. I’m a bot and I’m open source!
