Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The race to replace Redis

lwn.net

Post reply on HN
Print view View original

151–160 of 520 posts

Re: The race to replace Redis

#151

Post by BoorishBears »

busterarm wrote: post #118
BoorishBears wrote: post #73

Earlier quoted context omitted.

OpenSearch infuriates me to no end. It lacks so many improvements and advancements since the ancient version it was forked at, but because AWS already has an org's payments details, teams often refuse to look at Elasticsearch. Even basic things like autocompleting queries have been WIP for half a decade now: https://github.com/opendistro-for-elasticsearch/sample-code/... https://github.com/opensearch-project/OpenSear…

Teams should refuse to look at Elasticsearch. It's license is SSPL and they ship free and non-free features in the same binary. It's a ticking time bomb to run it in your company. Also you can just keep your data in postgres and use paradedb and stop having to deal with dramatically more expensive infrastructure and the JVM.

Ah yes, battle-tested Elasticsearch is a ticking time bomb for not wanting to get their lunch eaten by Jeff Bezos.

Just use this pre-V1 public beta software I stumbled upon instead.

Re: The race to replace Redis

#152

Post by photonthug »

lolinder wrote: post #117
klabb3 wrote: post #70

Why don’t we try to fix the “cannot be used for bezos yacht”-licenses instead of shunning the numerous companies of especially databases who want to do good in a meaningful way? Source available is good, better than proprietary which is what we get with aws, but still not enough. People are legitimately afraid of rug pulls, like sneaking in essential features into paid offerings. I think a lot of the skepticism comes…

> that you can slap on a project without a legal team The thing is, this kind of license is only really relevant to the kinds of projects that do have legal teams. If you're writing a hobby project you probably shouldn't waste time worrying about feeding the AWS machine, because the odds that you'll get noticed and used are tiny. Just pick GPL or MIT and be done with it. If you're participating in a large decentraliz…

> The thing is, this kind of license is only really relevant to the kinds of projects that do have legal teams

So you want to advocate that every future database / infrastructure company needs to burn part of their runway to hire lawyers to do the repetitive work of making sure they can both try to be open and try to continue to exist? Plus we, the users, get to try to decode reams of legalese instead of using a convenient three-letter handle for an industry standard, like GPL or MIT? This does not seem ideal..

Re: The race to replace Redis

#153

Post by dragonwriter »

klabb3 wrote: post #70

Why don’t we try to fix the “cannot be used for bezos yacht”-licenses instead of shunning the numerous companies of especially databases who want to do good in a meaningful way? Source available is good, better than proprietary which is what we get with aws, but still not enough. People are legitimately afraid of rug pulls, like sneaking in essential features into paid offerings. I think a lot of the skepticism comes…

> Source available is good, better than proprietary

“Source available” is a subcategory of proprietary, not “better than proprietary”.

> But my edgy take is that most people don’t really care about deep ideology

I think most people that orefer open source to proprietary software either care about the business benefits open-source provides over proprietary (including source-available) software or have an ideological affinity for Free Software, occasionally both.

Re: The race to replace Redis

#154

Post by thayne »

ajmurmann wrote: post #113
dehrmann wrote: post #93

Earlier quoted context omitted.

> re-selling your product” and not much more That's not what AWS is doing. AWS is selling management services. The fact that managed DBs are as popular as they are says this is a significant value add.

And that's also how DB companies try to monetize. So a hyperscaler offering this directly really undermines your entire business. In the past you could offer a Enterprise version with support, but with the move to the cloud that market is shrinking and Amazon is eating the new market themselves

Perhaps we need a different way to fund database development (not necessarily a single company monetizing it).

If the service you provide is hosting DBs, you are are at an inherent disadvantage competing with hosted db offerings form your potential customers' cloud provider. Even if your product is technically superior in every way, you are another entity they have to do business with (billing, support, contracts, security evaluations, etc.), which adds friction, and either you host on your own infrastructure, which means higher network latency, and network costs to get data to and from your customer's cloud, or you have hosting options that run inside all the major cloud providers, in any regions your customers use, which means you (or your customer) ends up paying the hyperscaler for the infrastructure, and you have the added complexity of having to know how to manage it on multiple cloud platforms. And there there is also the fact that it is much more difficult for you to build integration with the cloud's IAM or other services.

Basically, most cloud customers would rather use a service that is part of the cloud platform than from another provider. Ideally, instead of competing with the hyperscalers, they would sell some service to the hyperscalers that have the ir own hosted services. But I don't know how to get there.

As a brief sidenote, AFAICT this isn't what happened with the hashicorp license change, for them it seems like the pressure largely came from startups, not the big cloud companies.

Re: The race to replace Redis

#155

Post by pabs3 »

klabb3 wrote: post #119
aragilar wrote: post #95

Earlier quoted context omitted.

Define "fix". By definition you cannot have an open source licence which says "cannot be used for bezos yacht". Either you accept that, and don't rely on exclusivity for income (which really what the whole relicensing thing is about), or you don't open source your code (and accept that not being open source is a problem for some people). Open source + exclusivity for income is an unstable state, and really only works…

> Define "fix". It’s early. Everyone is confused. If I could define it, I would have provided a defintion. At this stage, it’s about acquiring requirements and looking at prior art. And being humble about the solution space. No? If you don’t think there’s any problem today, then argue that point. > By definition you cannot have an open source licence which says "cannot be used for bezos yacht". By definition by what…

> By definition by what definition?

By the "Open Source Definition":

https://opensource.org/osd/

Re: The race to replace Redis

#156

Post by llm_trw »

akoboldfrying wrote: post #84
llm_trw wrote: post #80

Earlier quoted context omitted.

The best idea I've come up with is a license which only grants the rights to a natural person to use the software otherwise it is identical to the MIT, GPL or AGPL, whatever your cup of tea is. If you're a corporation then you need to buy a license.

This could be an interesting idea, but how would this constrain incorporating the licensed software in a larger piece of software? Either as a library, or a component like a Docker image? Would it be "viral" in the sense that, if I want to publish software that internally uses a Docker container running software with such a license, my own software can be used only by natural persons?

Yes, you will have to publish under a license with the same clauses.

Not because you are distributing it, but because only natural persons can run the software.

Re: The race to replace Redis

#157

Post by fractalb »

crabmusket wrote: post #146
fractalb wrote: post #116

I feel copyleft licenses look more favourable at this point of time. What’s the value of more free/business friendly licenses if you can’t guarantee that the same license will apply for all the future releases? Looks more like a bait and switch policy.

Am I right in understanding that the relicensing was possible because of the CLA, not just because of the BSD license? Would a permissively licensed project that didn't use a CLA be vulnerable in the same way?

GPL mandates that all derived software must carry the same license. No need for CLA, as I understand it.

Re: The race to replace Redis

#158

Post by jillesvangurp »

jacooper wrote: post #32
west0n wrote: post #24

Neal Gompa opened a discussion on the Fedora development list, noting the license change and the need to remove Redis from Fedora. Gompa also raised the issue on openSUSE's Factory discussion list. After Docker was phased out, various distributions have adopted the compatible Podman as a replacement for Docker. It seems that a similar story is unfolding with Redis.

Docker was only phased out in red hat distros because they don't like it and want to push Podman. Others still have docker packaged in their repos.

docker-cli is still open source (Apache 2.0) and being distributed in most flavors of Linux. Docker the company does not own all the source code. But like redis they are free to build their own non open source products around this code base.

Re: The race to replace Redis

#159

Post by dragonwriter »

meowface wrote: post #150
lolinder wrote: post #117

Earlier quoted context omitted.

> that you can slap on a project without a legal team The thing is, this kind of license is only really relevant to the kinds of projects that do have legal teams. If you're writing a hobby project you probably shouldn't waste time worrying about feeding the AWS machine, because the odds that you'll get noticed and used are tiny. Just pick GPL or MIT and be done with it. If you're participating in a large decentraliz…

People need to make money somehow. Developers who spend years creating, maintaining, and continually improving an open source database (or other project) used by millions deserve compensation. This doesn't apply as much to Redis Labs since they swooped in much later, but the general principle of trying to monetize your project with source-available licenses doesn't feel unethical to me. You're right that it's probabl…

> This doesn't apply as much to Redis Labs since they swooped in much later, but the general principle of trying to monetize your project with source-available licenses doesn't feel unethical to me.

Yes, monetizing with a proprietary license, whether source available or not, doesn't seem unethical to most people outside of Free Software ideologues.

“The licensing model isn't unethical but competing ones are” isn’t why open source licenses became popular over proprietary (including source available) licenses, the fact that they commoditized the underlying software, enabled competing orojects evolved from the same codebase on essentially equal terms (which also allowed a competing project to fully replace the original if the original at some point failed the community) and, as hosted offerings became more common, the zero licensing friction for hosted solutions, that's what did it.

It does mean charging monopoly rents for a hosted service isn't a viable way to recover development costs and pay returns to VCs, but until fairly recently, no one was trying to do VC-backed startups around single open-source products with that as their whole business plan, and the arguments as to why that would be a bad idea were well developed by the mid-1990s

Re: The race to replace Redis

#160

Post by fractalb »

fractalb wrote: post #157
crabmusket wrote: post #146

Earlier quoted context omitted.

Am I right in understanding that the relicensing was possible because of the CLA, not just because of the BSD license? Would a permissively licensed project that didn't use a CLA be vulnerable in the same way?

GPL mandates that all derived software must carry the same license. No need for CLA, as I understand it.

I misunderstood your comment. Yes, CLA's make it possible to change the license. I guess CLA's won't work for GPL'd software.
Post reply on HN