They build their own custom nitro ASICs to integrate with KVM, for networking, storage, security, etc. That code doesn't make sense anywhere but AWS. They build things in ways that are tightly integrated with their internal systems and at a scale that few have experience with.
Take their new Envoy based service mesh for example, I really do hope that they make meaningful contributions back to the main project, but I also suspect they will doing lots of custom integration with their own hardware and internal network that will never make sense to support in an OSS project. Who knows, their Annapurna Labs divsion might even be building a custom card that includes Envoy functionality.
As to them being known for destroying OSS businesses I am drawing a blank there, can you provide some examples?
From what I recall DynamoDB is a somewhat hacky API layer over a bunch of MySQL clusters. I'm not sure anyone would, or should, attempt to use a non-managed version of it.
The recent MySQL info comes from this 2016 thread on DynamoDB storing empty strings: https://news.ycombinator.com/item?id=13170746
Confirmed by this comment specifically: https://news.ycombinator.com/item?id=13173927
The latest 2018 ReInvent deep dive on DynamoDB doesn't reveal anything but still fits if mysql is powering the storage nodes: https://www.youtube.com/watch?v=yvBR71D0nAQ
Releasing stuff isn't simple, the hosted services only work because of the rest of their infrastructure. Writing about the tech is usually the best way to spur clean open-source versions that we can all use.
But there is more to this as well. AWS is a huge ecosystem by itself, a lot their services intertwined with each other, simply open source DynamoDB isn't enough without open sourcing the networking/storage functionality that is probably provided by other services.
Open source takes work and the foresight to separate your custom shit from things genuinely helpful to the community. Amazon doesn't encourage or incentivize that kind of thinking.
That is exactly what AWS has done in releasing Firecracker[1]. Better late than never.
If they're going to compete on $1.28 shower hooks there's no proprietary SaaS with millions in MRR on AWS that will be too hard to copy.
To me, the fact that AWS is basing all their services OSS is a good thing for consumers. The thought of it all being proprietary is what would scare me.
I think amazon are unlikely to give up any competitive advantage they are not forced to give up.
I'm not sure I'm reading your statement correctly, but it sounds like you're saying that anything not developed under a copyleft license is proprietary.
If it was important for authors to receive contributions from derivative works, why would they have selected a license that doesn't require that?
I agree that your statement is correct.
That said, I agree with your argument, but found it rather funny that you just picked two AWS services you can indeed run locally.
[1]: https://docs.aws.amazon.com/amazondynamodb/latest/developerg...
I thought the point of permissive licenses over copyleft licenses is that you're explicitly granting companies the right to make money off your code without giving anything back.
If you want a quid pro quo, then you should have been using copyleft (or something else) all along.
Tim O'Reilly has a story (which I'll probably butcher) where after publishing the BSD manuals as books he was accused by someone at a conference of stealing the author's work. One of the authors stood up and said something like "Actually what Tim is doing is exactly what we want - spreading the material as widely as possible and making it as useful as possible. We knew exactly what we were doing choosing that license".
It's the same spirit in which MacOS being based on BSD is fine, even if not a single line of code is shared back. The question isn't "What did Apple share back to BSD programmers recently?", it's "Is the world a better place with MacOS in it as something people can choose to use?".
The same applies to Amazon. It simply wouldn't have been possible to build something like Amazon purely on proprietary code, running only proprietary OSes and services. Even if they had, it would have been a lot less interesting. I think it's pretty clear that the world is a better place with services like AWS available. As long as they follow the licenses of the software they use, there isn't a problem, but if some licensors disagree they can always change their license.
https://writing.kemitchell.com/2018/10/24/How-to-Speak-Copyl...
Many either don't understand or don't want to take on any risk associated with some kinds of licensing or barriers in certain situations. Right now I've had to work around calling a GPL executable (pngquant), simply because we don't want to have to require it as a separate install. So switched to an alternative implementation.
It just depends on where you are, where your business is, and what you're working towards/against. It's hard to find a balance. And it will take some time to flush out. MongoDB and Redis come to mind here. They have to make some money to pay their bills and developers, also profit isn't a bad word. At the same time, huge cloud providers are offering services using their software and the devs make nothing. I get both sides here. There's no perfect solution here.
When the Spectre/Meltdown hit, Amazon also open sourced the mitigation used for the AWS version of Xen ( http://xenbits.xen.org/xsa/xsa254/README.vixen )
If you search the Linux git commits for people with @amazon emails, you'll find contributions, bug fixes, etc.
I don't think it's fair to say that barely anything is contributed back.
It was announced last week during reinvent.