And Google is probably ecstatic. It's disrupting the AWS monopoly that was developing, tons of outside developers are improving GKE for them, and it's running a textbook "commoditize your complements" play.
Compare that with Android, which Google is successfully doing the EEE strategy on. Despite widespread usage, it has few outside contributors, so Google is bearing all of the costs of it being open source without reaping few of the benefits.
But that's Google's fault. It's hard to build and contribute to Android. Development is done privately and thrown over the wall.
OTOH, it's easy to contribute to Kubernetes, and it's done in the open.
So a few years ago I created an open source community in gaming (which coincidentally, the author of this blog post was a part of early on). We had a Github organization, an IRC channel, etc. It was fun.
At some point I grew part of the community into a company, which today is growing and thriving. We're rooted in open source. I'm rooted in open source (Nearly all my work has been public and on Github for the past decade).
Today we still have nearly all our work up and public on Github but get next to zero external contributions. Because contributing as someone external to the company is hard.
Unless your actual goal as the company is to produce free & open source software, it's going to be really, really hard to make it possible for external actors to contribute to the core products as smoothly as for the company's employees. There's loads of things that pile up, like reusing internal CI systems for testing, requiring API keys and passwords for access to resources that are required or highly useful for development (eg. testing a redshift environment), and so on. All those things are fixable, but take a huge amount of time to fix, and why would you fix it if nearly nobody external to the company is contributing anyway? It's not like making the experience easy is a guaranteed win either.
Things like product design, project management, etc... it's not obvious how to make those open processes. It's not like you can do part time project management contributions to any random open source project out there. And when the processes aren't open, the barrier of entry to contribution is that much higher.
And quite frankly, certain people (who will recognize themselves) make it extremely unappealing to spend a lot of efforts on keeping the entire project open. The kind of people who will complain about every little detail of how you're trying to open up the project, until its name is prefixed with GNU/. "All or nothing" is an atrocious approach to open source and actively harms the ecosystem.
Point is, don't blame Google for making Android and Chromium hard to contribute to. It's not like it's on purpose, it just takes a huge investment for it to be possible. I mean fuck, Firefox isn't exactly easy to contribute to, and Mozilla is one of the most open companies out there, and in many ways a role model of how an open source company should be run.
In open source, there's invariably an "in-track" and an "out-track." The in-track consists of people who will get their concerns pretty quickly addressed and can make their patches known and reviewed fairly quickly. The out-track is those who can't do that.
The strengths and weaknesses of open-source models is in how easy it is to make it into the in-track, and how stark the in/out dichotomy is. My experience with Mozilla is that they do a fairly good job at speeding people from out-track to in-track (although I suppose it depends on what areas you work in--Thunderbird was ridiculously easily, but I can imagine Gecko or SpiderMonkey is going to be much harder), but the in/out dichotomy remains somewhat stark. I've seen other places, like LLVM, where the in-track is much harder to break into, but the in/out dichotomy is not so stark.
I have no idea why this didn't happen for Android. Samsung & LG have spent a fortune on Tizen & WebOS, and would have been a lot better off with a truly open Android. Whether that's their fault or Google's is hard to tell, but I suspect that answer is "both".
It’s a pretty hackable project.
They probably won't.
Further, Google has been taking the Google bits out of Kubernetes. They declined to call their own offering "Google Kubernetes Engine" officially until the playing field was level and they had a certification program. Amazon, Azure, DigitalOcean and others are implementing their own managed Kubernetes solutions.
Google doesn't have to cheat here. Speaking as someone dealing with Google Cloud directly, they are mainly hitting on the point that Google Cloud offers better network and disk performance, in a more flexible manner, for significantly less money. And honest to God they're right. The CPU performance is worse in some cases, but we're saving money on workloads so far.
On the flip side, as long time AWS users, the Kubernetes experience has been pretty good since Kubernetes 1.3. New features typically support AWS pretty early in, with the main exception being Ingress controllers (for good reason: ALBs weren't even released when that became a thing.)
Google is one of those companies big enough to make it easy to feel like they have evil ulterior motives, but I'll be honest: in my dealings with them, I've felt that the opposite is true and generally the intentions are good.
I have more confidence in Amazon and Microsoft.
Google sticks with things that are core. Google reader was obviously not core.
Glasses moved to enterprise and continue to be used in that setting.
I'm not saying it's better than other offerings - It's not all around - but the building blocks are really nice.
There is a lot to be said about choosing the default. I've been a .Net developer for over a decade, but when I started looking into Lambda for instance and saw that it looks like most people were using Node and to a lesser extent Python, I begrudgingly started learning Python. In hindsight, I'm glad I did. I'm actually starting to like Python and I usually hate scripting languages.
Besides, for my own selfish reasons, why would I spend time on the cloud platform that is the least marketable? Companies are asking for AWS and to a lesser extent Azure experience. Very few are asking for GCP experience.
If everyone is going one way, it's not necessarily a bad decision to take a slightly different path. I've been able to build a career and a business on the alternate path.
During those days UNIX based systems were just yet another option, mostly on government and university departments.
I learned UNIX thanks to our OS teacher carrying around a single PC tower (remember those?) with Xenix on it, where we would take 30m turns to try out our C programs, prepared in advance with Turbo C on MS-DOS.
80's and early 90's, before Linux and BSD actually took off as a free way (free beer) to get access to UNIX.
They are giving away everything they learned with Borg and you somehow take it as a negative? Really?
Do think from a business standpoint not sure why Google would give up the K8s brand but it is good for everyone.
I really worry that Google will stop opens sourcing and more important sharing so much through papers when people b*tch so much.
Starting to be hard to see why they even do it any longer. Who else have a huge competitive advantage with something like Borg and just gives it away to the industry?
But I am someone super curious and just probably spend more time on
https://research.google.com/pubs/papers.html
Read most of the paper. But then what they presented at NIPS. My favorite paper right now is
https://arxiv.org/abs/1712.01208
Bigger example is how much they give away in terms of AI/ML.
You think they are publishing all these papers and open-sourcing TensorFlow for the good of the community solely?
No... if the entire ML community publishes papers with TensorFlow as code, then now Google can sell that many more TPUs and that much more access to their cloud systems to run the models...
I was shocked they gave away the K8s brand. Now that was ballsy and surprised necessary.
Because with Borg they only had an advantage with their own products' operations. That's nice, in that they are able to maintain a decent amount of technical superiority and have their product reliability be leading or tied for first, but it's not a clear line between this and them generating more revenue. They want their cloud offering to compete with AWS and Azure, which is why they want K8s to gain mindshare. GCE has huge potential for growth and significant switching costs; meaning they need to get new products on it sooner rather than later, and developer mindshare is the way to do that.
They were tired of writing something up as a paper and then someone implemented as OSS and then companies and engineers knew the OSS implementation instead of their own.
So instead do a OSS version and then drive what everyone learns.
I found this a bit funny. One solution would have been to not write the papers. But so glad Google does. Spend a lot of time on their research site.
The original papers on Map/Reduce, Bigtable, Dremel, Dapper, Spanner, directly spawned a bunch of successful open source implementation projects as well as companies.
But with Google-run (or other company-run) open-source projects, it's a completely different story. Want to be a TensorFlow maintainer? Great, sure, you can contribute, but you'll never get to make any decisions about the direction of the project itself unless you're inside Google.
To me it's a disappointing semi-abuse of the term "open source"... it doesn't mean what I think it should...
The larger and more important a project, the more bureaucratic it will be, and the more people you’ll be competitive with for attention.
...you could always find other people who think the same way and create a fork.
This isn't unlike any other open source project ran by people who don't want to consider all input from others. Not all suggestions are worth serious consideration, after all. Open source doesn't mean I'm beholden to your every whim just because you've contributed. That's what forks are for.
Don't want to move to California? Can't pass the interview? Salary too low in the offer? There could be lots of reasons.
I agree with your other points, but what I'm saying is not exactly that. What I'm saying is that TensorFlow is not a community-led project in the way that other open-source projects are. It's Google-led, and if you don't like it, you can go away or fork it (...and if you fork it, it's unlikely you'll get many users given the Google marketing machine).
(The same complaints have been and are being regularly made against Red Hat, and also Fedora, given that Fedora is basically "experimental" Red Hat.)
Consider if Tensorflow were half as useful as it is today if it was community-led and it was led by a shitty community.
Would you have preferred that those developers reverse engineer something closed source Google developed instead? Libpurple has advanced messaging not at all, only been great for interoperability.
Google talk / now hangouts once was an XMPP pusher. XMPP is decentralized. As with mail servers you can advertise your own XMPP servers, see https://wiki.xmpp.org/web/SRV_Records , you can even publish a per-user DNS record with an OTR key.
I shut down my XMPP endpoint a few monthes ago because Google had stopped supporting XMPP bridging to hangouts. Now that everybody uses hangouts (coders) or slack (teams/corporates) or whatsapp/messenger/skype (everyone else) there is no incentive to support it anymore (I speak now only IRC with coders)
They gave their support and a lot of traction to the protocol, and removed it on a second step (once advocates of XMPP, _yes, like me_ said to everyone, "yes you should prefer gmail+talk, they are supporting open standards")
Looks like an approaching-with-good-intentions and then disrupting strategy to me. Even if you disagree on the responsibility of XMPP rise and fall, they are the bull in a china shop : don't drink the kool aid, you'll end up broken.
Google backstabbed XMPP, sure, but it's healing pretty well. I don't see how any time spend by developers on XMPP is wasted just because Google backed off. They took their userbase with them, sure, but if you want modern, open, federated IM network, XMPP is a sane choice. The only proper alternative these days is Matrix, but that's more of an IRC replacement.
The main externality of this open source EEE strategy is developer opportunity cost. When google swept into the chat market, they used their size and first mover advantage to push an open standard, XMPP. So developers who might have worked on other projects/standards naturally gravitated toward XMPP and contributed their time there. If those developers had known the future of XMPP, they likely would not have worked on it. They could have invested time in a competing, more open project instead.
This to me is the real danger. It’s bad enough that google gets to benefit from thousands of free contributors to its “open source” code. It’s even worse that they are taking those developers away from other projects, so that when google moves to the third phase of EEE, there are no viable alternatives.