Amazon slams media for not saying nice things about AWS, denies it strip-mines open-source code for huge profits Stung by an article mulling Amazon Web Services' market dominance on Monday, AWS VP Andi Gutmans fired back, complaining the reporter ignored flattering comments from AWS partners – and that "AWS is 'strip-mining' open source is silly and off-base." "The journalist largely ignores the many positive comments he got from …
Well duh, they're your partners. They're not going to trash you, they need you. The journalist may have had an agenda, but if the truth hurts it's not his fault.
Besides, what's the problem ? It's not like your customers are suddenly going to disappear if you admit to "strip-mining" open source code. Many people have been warning about using someone else's server and that hasn't stopped them from using yours, so your use of open source is certainly not going to stop them either.
Amazon (and not only Amazon) is terrified it can't no longer exploit others' code to pump and run its offerings.
AWS, Facebook and Google wouldn't exists - or at least wouldn't be so profitable, without being able to use and run a lot of code for free, and on a lot of systems.
The FOSS people ignored Google and Facebook partly because they couldn't see the direct connection between "free software" and the enormous revenue streams, and partly because that companies propaganda of being "FOSS friendly" worked - as they give them some breadcrumbs to keep them happy (while keeping all the relevant, money making code proprietary - as it's never "distributed").
With Amazon, the connection is much more direct and visible - Amazon does rent the usage of that very FOSS software, or makes copiesforks whenever the license is changed trying to stop them. If more and more companies start to block Amazon, AWS has to pay for licenses, or develop that kind of services from scratch - which is far more expensive - and could make the whole cloud business far less competitive.
Of course customers can also buy their own licenses and install the software on cloud systems, but that would be money AWS & C. would never see.
While of late it's been difficult to defend Google on many topics, I think it's reasonable to suggest that they make a tremendous contribution to open source software.
They've created and released many libraries, quite apart from the people working for Google sharing updates and changes to existing projects.
Amazon and Facebook do also make contributions to open source projects and release libraries under open source licences. Whether Amazon contribute enough would be a reasonable question but also one with a deeply subjective evaluation criteria.
The top five entries on the bottom chart in this article feel relevant: https://thenewstack.io/week-numbers-data-lots/
Amazon, Facebook and Google are all also sponsors of the Apache Foundation (and likely other initiatives) so they're certainly making at least some form of financial contribution too.
Sure - for the part they need and it's not their main business. But they took and take much more than they give back. None of Google critical and money making code is open source - still Google won't be able to operate at the scale it runs without a lot of free code it doesn't have to pay licenses for.
But keep on thinking they are good because they give you some web libraries.... Facebook does exactly the same - and just because it actually fully supports their business - gathering people data and selling ads.
"Amazon (and not only Amazon) is terrified it can't no longer exploit others' code to pump and run its offerings."
I'm going to disagree - I'm not sure it is AWS that is terrified with the current state of play. And this downplays the effects of the non-US providers that benefit more than those doing significant development in-house.
MongoDB/Elastic/Redis were expecting their market to be the enterprise/third-party users to provide their sales with the big cloud providers using their own or equivalent products long term. The market hasn't developed that way with a significant proportion of the enterprise market for these tools going straight to the large cloud US providers that offer additional features targeted at scalability and reliability.
Is it "fair"? For Redis which is largely original content, probably not but for MongoDB and Elastic which were building on existing open source products, they relied on the open source model to commoditise and develop/test their products while relying on VC cash. And in the process, a competitor has taken advantage of their openness to release products that more closely match their requirements. They took a risk with relying on others to help them with development and testing and it hasn't provided the returns they have hoped for with their current offerings.
While it harm open source in the long term? I don't believe so - this has been happening for decades now and if the cloud providers can't develop their products in ways developers/end users require, we will see forks and better alternatives produced over time. It may alter the way VC's fund startups or require new startup business models but I'm not sure their will be too many crocodile tears for the VC's.
"AWS, Facebook and Google wouldn't exists - or at least wouldn't be so profitable, without being able to use and run a lot of code for free, and on a lot of systems."
AWS & Facebook, maybe. But I just looked at the 2017 kernel report. Google is 10th in company contributions for 4.8to 4.13, seventh if you strip out "none", "unknown" and "consultants". Only 4 changes less than SUSE and 3x the number by Canonical.
Just because actually Google does distribute Linux - so it's forced to make far more changes public.
And why should you strip unnamed contributors - which actually looks to write more code than the mighty Google?
I think the point you and everyone else misses is that AWS does nothing wrong.
Has anyone actually bothered to read the Apache license?
It says that you can use and take the code and do whatever you want with it. (With some very liberal requirements.) If you donate code... you have to indemnify Apache if they get sued. So if you work for ... lets say IBM and even if you wrote some code on your own time on a non work project... IBM could still sue Apache if the code had any IP value to IBM. (Even though they donate code/projects to open source. (Cloudscapes Silver Cloud which became JavaDB ... or Eclipse which was based on Rational code)
It seems the rest of the world is waking up to what is happening in the open source world. and why its not the panacea everyone thinks it is...
Sorry, not that I'm defending AWS, Google or half a dozen other major companies who exploit the system... just that this isn't really news. AWS mimics the tools people use on their infrastructure so that you can ditch those companies and use their internal source controlled copies.
Google, Amazon need to be broken up but the argument is a complicated one.
You might like to think about the difference between right and wrong, and legal and illegal. Open Source economics tried on a community both using and contributing. For one of the largest beneficiaries of Open Source to give virtually nothing back to the community, and to seek to remove the companies that actually do contribute, then it's very dangerous for the wider open source community.
There's a big difference: Amazon can stop you using its servers as soon as you don't pay, companies distributing open source code don't have this easy option to stop Amazon to use their software when it's distributed with common open source licenses.
Yet, for customers not knowing if the third party software and services they're using will be still available tomorrow or not, or if they have to obtain them separately, plus install, configure, and manage them themselves can be a big issue - which can impact their decision about moving workloads to AWS or not.
Plus, there's the "bad guy" image Amazon doesn't like - just look at how much they also work to fend off any critics of their goods retail service, and its workers exploitation.
The open source companies could have been entirely proprietary if they wanted. They chose open source either because a) the initial IP was released by another company, even if the founders were involved it was paid by someone else, and/or b) they knew that adoption depended in part on being free. You can tell this is the case when the open source company, such as Mongo, retains the license rights from outside contributors -- they want to themselves to control the destiny rather than the community. The implicit bargain they are making is "hey devs, adopt our tech for free with no risk to see if it works for you, and later when you go into production make the operations guys pay because we are best placed to support". This bargain assumes they will be the best choice for support, but what if the cloud is a better support option? Yes it breaks the bargain, but it breaks it on both sides. This bargain requires the open source companies to provide better support than the cloud providers. If they flip it to proprietary, there's a different bargain required with the developers as the risk profile changes. It's entirely possible to do that -- look at Snowflake, no open source. Others are making a go at it is Databricks, their cloud implementation is proprietary but they are competing at the operations level, maintaining the open source bargain. These other companies need to decide where they stand and compete at the operations level, otherwise sooner or later the cloud will kill them even if it's not AWS, it will be Google or Microsoft or Alibaba or someone else.
In this case the "operation guys" is AWS - just it's big enough it can support itself.
Sure, they could have gone closed source, but lately looking like IBM, Microsoft, or - god forbid - Oracle, is a no-no among cool developers, all the code has to be free and published exclusively on GitHub (which eventually became Microsoft, money has to come from somewhere eventually...), or you're frowned upon.
It's quite clear that even the GPL can't help in this case - it will be interesting to see how it will impact the next generation of code, as the FOSS business model may have lost one of the few revenue chances it had.
"Cloud is just other people's infrastructure."
AWS is coming up to its 14th year of operation - do you think this is still relevant?
Cloud has presented and demonstrated what it can and can't do.
Disclaimer: my current organisation doesn't use cloud for core business systems but we are dependent on other peoples offices, data centres, WAN links, Internet connections, electricity etc
Journo writes largely-pointless attack on Amazon. OK, not entirely pointless, it's journo's job to produce columns.
Amazon person writes similarly-pointless rebuff. See above.
Reg hack writes even more pointless attack on Amazon rebuff spot the pattern?
Commentard vanishes up his own pointlessness continued on page 94.
You'e complaining of that which makes the world go 'round, and has for a very long time.
OK, fine, no problem. But far more useful is to provide some other way to keep the wheels turning, one that we can all adopt - and keep those kids fed.
You do know it can be quite entertaining, right?
How long has IBM been running Linux VMs on its big iron? Not just since it bought Red Hat. Any bets on the revenue split there?
How much more did Facebook make off the LAMP stack than the developers of Linux, Apache, MySQL, and PHP? (Yes, I know they aren't pure LAMP, and haven't been for a long time.)
What is the role of OSS in Google?
Yes, Amazon definitely falls in the "view with concern" category. But I thought the story could have been more nuanced.
I suspect similar to this taken from Commoditize Your Complement which is a summary of this larger article from Joel on Software
"Joel Spolsky in 2002 identified a major pattern in technology business & economics: the pattern of “commoditizing your complement”, an alternative to vertical integration, where companies seek to secure a chokepoint or quasi-monopoly in products composed of many necessary & sufficient layers by dominating one layer while fostering so much competition in another layer above or below its layer that no competing monopolist can emerge, prices are driven down to marginal costs elsewhere in the stack, total price drops & increases demand, and the majority of the consumer surplus of the final product can be diverted to the quasi-monopolist. A classic example is the commodification of PC hardware by the Microsoft OS monopoly, to the detriment of IBM & benefit of MS."
You have huge companies in place that have a massive marketing name, if they want to take you OS code and repackage it in their own products there is naff all you can do about it. The OS label just put a neon light saying "Free Stuff" above your project. Who is going to go with "Startup Inc" when the exact same stuff is available under "Mega Corp Inc"?
So, the story seems to be companies make software and offer it for free, in the hope that they might be able to make some money off the back of it at some point.
Other companies take the software and use it to make money.
Company is sore that their business model isn't paying off. Whines at successful companies with a workable business model.
It seems to me that the real problem is that companies betting their bucks on such a tentative business model should suck it up when it doesn't pan out the way they hoped. It happens to a lot of companies. Perhaps they should be competing with AWS or Azure instead of whining about how hard they have it?
The Document Foundation and Apache seem to be doing alright.
Those that don't depend on cloud services are doing okay and can carve out a niche, those that do are just having their work ripped off by the cloud services oligopoly. How did the cloud services oligopoly get where they are today? By open source.
The business model did pay off, it doesn't now. As a start up, your established competition has made billions innovating using open source as a base to start with and has reached that point where it doesn't need to innovate any more. Perhaps the internal culture won't even allow innovation any more, maybe it will only allow money to be invested in things it 'knows' will be successful.
Those megacorps which are already there have billions to invest in staying on top of the pile. They do that by copying and pasting open source and buying up promising apps or services by start-ups and shutting them down.
Google has announced they may bail from cloud services in a couple of years if they're not successful (this 'burning platform' announcement may even help that happen). If they can't do it, nobody can.
Is it moi, or do any others think companies like MDB should not whine because they were able to launch an IPO on the backs of a scheme to support software that is open.
Like this business model was ever going to make a profit??? It's OPEN SOURCE and, therefore, fair game for AWS to flavour up its own service and sell it to its hosted customers cheaper than Mongo sells their stuff for.
Companies like MDB raise all this venture capital. The venture guys get to the banks to underwrite the IPO. The banks make a killing in fees. The stock launches and fund managers who don't really get what MDB even does suck up shares, driving up the price and it looks swell to everyone so long as the stock keeps going up and it's not in your retirement account when it crashes down.
It's a bloody racket making insiders rich. AWS has the perfect right to destroy them if they want. They own the damn data centres.
However childish AWS's responses was, I do feel for them in this case (and it's a lot less childish than the continuous digs at AWS in Elastic's blog posts), partly because the article they were complaining about looked like it came straight out of elastic's marketing department, it wasn't journalism in any form.
Elastic have been a particular poor open source citizen in this case. Firstly, having attracted contributions from many (including me) into the open source elasticsearch code, Shay has now gone back on his statement that elasticsearch would never be 'open core'. It's now a commercial product, the fact that a small amount of code is still open source doesn't make it an open source product. Secondly, they've merged open and closed source into the same repo which shows the direction they're headed in, and the extent to which they (don't) value their community.
When Shay chose the Apache licence for elasticsearch, he did so because he wanted people to use it commercially (otherwise he would have chosen a different licence...) because he needed the contributions that those companies (including AWS) would make to his code. He's now making quite a lot of money out of some code which has been contributed to by those same commercial organisations. In addition, elasticsearch is heavily dependent on Apache Lucene - in the latter years the main contributor to lucene was employed by IBM (to work on lucene full time iirc). So the money that Shay is making from elasticsearch now is partially driven by code that IBM spent money creating. It certainly feels like a contradiction to say it's ok for elastic to use code paid for by other companies, but then for some reason AWS isn't allowed to use elastic's code.
I'm pretty sure that the majority of elastic's complaint is simply that AWS are competing against elastic's own search-as-a-service, but they're trying to use allegations of poor open source practice to cover that up.
If Elastic don't want people to use their code commercially, they should use a licence that forbids it. This childish anti-competitiveness is helping no one except the lawyers.
/rant
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
