Docker Is In Deep Trouble (zdnet.com) 141
An anonymous reader quotes a report from ZDNet: Docker, the technology, is the poster child for containers. But it appears Docker, the business, is in trouble. In a leaked memo, Docker CEO Rob Bearden praised workers -- despite the "uncertainty [which] brings with it significant challenges" and "persevering in spite of the lack of clarity we've had these past few weeks." Lack of clarity about what? Sources close to the company say it's simple: Docker needs more money.
Indeed, Bearden opened by saying: "We have been engaging with investors to secure more financing to continue to execute on our strategy. I wanted to share a quick update on where we stand. We are currently in active negotiations with two investors and are working through final terms. We should be able to provide you a more complete update within the next couple of weeks." Docker has already raised $272.9 million, but the company hasn't been profitable. It's venture-capitalist supporters -- ME Cloud Ventures, Benchmark, Coatue Management, Goldman Sachs, and Greylock Partners -- which have seen it through Series E financing, can't be happy, that after almost six-years, Docker still isn't close to an IPO. While the previous CEO, Steve Singh, promised in May 2019 that Docker would be cash-flow positive by the end of this fiscal year, that appears not to have been the case. Otherwise, Docker wouldn't need to seek additional capital. ZDNet's Steven J. Vaughan-Nichols says the reason has to do with Docker's lack of a viable business plan.
"That's in part because Docker had hoped to make container orchestration, with Docker Swarm, its profit center," writes Vaughan-Nichols. "Then along came Kubernetes, and that was the end of that. Kubernetes has become the container orchestration of choice, leaving little room for others. And, indeed, Docker has adopted Kubernetes as well."
Indeed, Bearden opened by saying: "We have been engaging with investors to secure more financing to continue to execute on our strategy. I wanted to share a quick update on where we stand. We are currently in active negotiations with two investors and are working through final terms. We should be able to provide you a more complete update within the next couple of weeks." Docker has already raised $272.9 million, but the company hasn't been profitable. It's venture-capitalist supporters -- ME Cloud Ventures, Benchmark, Coatue Management, Goldman Sachs, and Greylock Partners -- which have seen it through Series E financing, can't be happy, that after almost six-years, Docker still isn't close to an IPO. While the previous CEO, Steve Singh, promised in May 2019 that Docker would be cash-flow positive by the end of this fiscal year, that appears not to have been the case. Otherwise, Docker wouldn't need to seek additional capital. ZDNet's Steven J. Vaughan-Nichols says the reason has to do with Docker's lack of a viable business plan.
"That's in part because Docker had hoped to make container orchestration, with Docker Swarm, its profit center," writes Vaughan-Nichols. "Then along came Kubernetes, and that was the end of that. Kubernetes has become the container orchestration of choice, leaving little room for others. And, indeed, Docker has adopted Kubernetes as well."
Not a problem (Score:2, Funny)
Pay the VCs back in Bitcoin.
Re: (Score:2)
Serious question- how and where would you sell/convert Bitcoins into actual dollars when you're talking about a sum of $250 million or so? Is there a bank or exchange that would take an amount like that?
Re: (Score:3)
Re: (Score:2)
With that amount, it's hard to do even if you are buying a heavily traded stock.
Okay, so let's say I had $100,000 in Bitcoin...where could I go to exchange it for actual dollars?
Ask Softbank for a big bag of $$CASH$$ (Score:2)
Since they blindly invest in any startup like WeWork and Uber, they must have some more money laying around I think.
Re: (Score:2)
Do they have a bigger-than-life CEO? (Score:2)
But their CEO seems rather dull ....
Re: (Score:2)
That money dried up, and assumed future investors pulled out.
Softbank's VC bad choices are over.
Re: (Score:2)
Nah, they just went into the pet care business.
No way (Score:2)
Re: (Score:2)
Cargo shorts forever !!!!
Google or Amazon should buy them. (Score:4, Interesting)
That would make total sense.
Re: (Score:2)
What is left to buy? Or even to steal?
Re: (Score:2)
Since Google basically owns the development of Kubernetes (Now the #1 deployment tool for Docker containers), it kind of makes sense for them to own the future of Docker.
Re: (Score:2)
Ni!
Re:Google or Amazon should buy them. (Score:5, Interesting)
I was thinking Microsoft. Microsoft distributes SQL Server for linux over Docker so it should be in their best interest to keep Docker afloat.
Amazon, on the other hand, probably already forked Docker for their own purposes and wouldn't care if it curls up and dies.
Re: (Score:2)
Re:Google or Amazon should buy them. (Score:5, Interesting)
No, Microsoft distributes SQL Server for Linux via containers - you can actually run the container on several of the different containerisation systems out there, including Dockers containerd...
People seem to think that Docker is "it" in the container space, and they really are not - there are very few incompatible standards for container formats, and you can run those container formats on most of the containerisation runtimes out there (containerd, rkt, mesos, LXC et al).
Docker at this stage in the game is a management system around containerisation - Swarm has already lost out to Kubernetes as the go-to platform for large platforms (despite Swarm being significantly simpler), and Kubernetes has built in support for swapping out the container runtimes. Docker has the Docker Enterprise Container Platform, which implements Kubernetes and a whole host of supporting infrastructure, and is where they want to make their money - they can't make money from plain old "Docker" because there are already decent free alternatives.
Re: (Score:2)
Too much power / too many companies (Score:3)
there are too many companies out thete relying entirely upon Docker for their containerisarion.
Amazon, Google or Microsoft buying them would be putting too much control in the hands of a single conpany: it's going to be the entire "Microsoft buys Github" scandal playing over again (with this time the role of GitLab played by LXC or Singularity).
The best solution would be to follow the same model as Linux or Mesa and have the big players that have interests in the tech having developers working on the tech w
F/OSS Advocates (Score:3)
For similar reasons, Ansible Tower will soon change or be forgotten in favor of AWX.
And none of these F/OSS shops provide the Cadillac support infrastructure to differentiate. At the end of the day, any company leveraging them still needs their own in house experts.
Re:F/OSS Advocates (Score:4)
The real problem is that companies that want to buy support for containerization are choosing RedHat and IBM. Er, I mean, IBM.
Re: (Score:2)
And so, when faced with a one-trick pony, go with the horse breeder.
The previous docker management regimes had some really crazy policies.... pushing mesos and not getting in front of kubernetes. They let their libraries get infected. Didn't have helpful outreach.
Indeed few technical evangelists really drummed up warmth for docker-the-org.
It could be turned around, but it's my guess they'll be subsumed into other groups or mutate like openstack. Openstack had great evangelists. And it proves that orchestrat
Re: (Score:2)
The problem with that is that most of these entities never had a marketable, sellable product before involving the community. They had a barely sketched out idea before getting contributions. They couldn't sell shit. So for most of them, the "develop a salable product before release" strategy isn't viable, let alone a good idea.
The follow up is of course that most of them don't actually have any reason to exist. We only need a small handful of container management systems. Most of them have no justification
Docker is a Company? (Score:5, Insightful)
I had no idea Docker was a company. I though it was the hip, new, FOSSy way of paravirtualizing shit.
Who in the hell though they could make money off of this?
Re: (Score:2)
I thought this was about khaki pants for sure.
Re: (Score:3)
Who in the hell though they could make money off of this?
Err they DO make money of this, as do the many other companies in this space. Docker specifically is just not making enough money due to competition. Seriously dude have you missed the whole "cloud" thing that this plays into?
new, FOSSy way of paravirtualizing shit
You could have just lead with "I don't understand what I'm talking about but I feel like talking anyway".
I find Pod Manager tool (podman) to be better (Score:5, Interesting)
I'm currently using it on a small scale on Fedora but I know of a couple of large projects planning the move it's out of beta for Redhat and CentOS [redhat.com]
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
I'm very interested in the daemonless part. I read the homepage but could not figure out this part:
How do I build a container and does that build save to an .ISO file or something?
if someone could tell me what docker is or does (Score:3)
I might be able to help, but I'm a technology dinosaur, I still like saving files locally!
Re:if someone could tell me what docker is or does (Score:5, Funny)
You know what they say: If you're not wearing Dockers, you're just wearing pants.
Re:if someone could tell me what docker is or does (Score:5, Insightful)
Say you're a shitty developer and you write a program that needs a dozen obscure libraries that have even more dependencies on top. If anyone else wants to compile your program its going to be a royal pain in the ass trying to locate all the pieces and recreate your process. Instead you make a docker container. A portable filesystem of your work that other people can mount with a click. But hey you're totally trustworthy and docker could never have any security problems...
Re: (Score:2)
I am pretty ignorant, bit isn't the point of docker to put things even more local (to the program rather than the system)?
Re: (Score:2)
Oh ... you mean a zip file?
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
It's better than that. Not only can you statically link the libraries, but you can hardcode the entire environment in your program.
shitty developer without knowing it (Score:5, Insightful)
The purpose of Docker is not to make it easier to compile with obscure libraries. A good developer would know that.
One of the main advantages of Docker is to combat “Works on my computer”. Being able to run exact code on all environments (local, test, acceptance, prod) is good. Not having to think about version conflicts between any dependencies running on any server is good. Easy scaling is good. Orchestration framework like Kubernetes is good.
Stating that Docker is for shitty developers, here on Slashdot, and getting upvoted for it, is just sad.
I highly recommend to learn about Docker and Kubernetes before dismissing it.
Re: (Score:3)
It's sad that someone would even write a comment like yours on slashdot, even more so that it would get modded up.
The "works on my computer" problem exists specifically because of crappy programmers, or at minimum, crappy maintainers.
People misuse APIs and then they are confused because things don't work later with a new version of a library. Or maintainers link to the major version when a program needs a specific minor version of a library to function... Although that comes back to crappy developer, who ha
Re: (Score:2)
I fail to see what is sad with me replying to that.
Your comment do have a point, but it’s still aimed at borderline misuse of containers.
You fail to see the security benefits of containers. Being able to remove a container and replace it with a new (patched) version without being worried about conflicts or dependencies is gold. Patching any system is a
Re: (Score:2)
You fail to see the security benefits of containers.
You fail to see that the emperor is naked.
Re: (Score:2)
Re: (Score:2)
The "works on my computer" problem exists specifically because of crappy programmers, or at minimum, crappy maintainers.
No. It exists due to variance in the Linux ecosystem. This problem is only combated by holding back versions or containerising, the former being a massive problem for end users. There are of course workarounds, some of which are able to fuck up your apt repo in ways that will ensure you can never upgrade your system again without doing a fresh reinstall.
Seriously are you a Windows 10 fan? You shit on Windows all the time, but right now you're acting like you've never even used Linux.
Re: (Score:2)
The "works on my computer" problem exists specifically because of crappy programmers, or at minimum, crappy maintainers.
This may be news to you, but companies hire crappy programmers all of the time and you are still expected to work with them.
I'm a programmer/admin that does Devops these days, the developers we hire aren't necessarily expected to be proficient in the entire technology stack, but they do need to run the entire stack on their laptops. Rather than spend the time teaching them all how to install and configure things like MongoDB or RabbitMQ and then deal with the all of the follow-up questions/requests I can
Re: (Score:2)
Re: shitty developer without knowing it (Score:2)
"One of the main advantages of Docker is to combat âoeWorks on my computerâ. Being able to run exact code on all environments (local, test, acceptance, prod) is good."
Oh my god man, this has been solved a hundred times over, on every shipping OS. https://en.m.wikipedia.org/wik... [wikipedia.org]
How about this, ship a zip file, or an rpm, or msi, or dmg, or dpkg whatever for minimal system integration, and then get ready for this... !!DON'T!! ASK OPS TO RECURSIVELY RENAME CONFIG FILES FROM PROD-*.XML TO *.X
Re: (Score:2)
Your caps rant, is all to true. Having run into in many companies. Even some of the leet's whee everything is in GIT and positions are DevOPS/SRE or full stack developers.
It's why there is a rift between developers and system admins.
Re: (Score:2)
You're talking about a very simple environment of a single server with some software packages installed along with some configuration files.
We have a grid compute system with a web frontend, messaging middleware and a database backend. Developers need to run all of these services on their laptops in order to do their development work. Docker allows us to put all of these services together in one place for the developers to use which allows them to focus on their work instead of having to configure all of
Re: (Score:2)
There might be some overlap, but they are really very different.
There are systems out there where a zip file and vi is the right tool for the job. There are also systems out there where they are not the right tool for the job.
We clearly have different knowledge, experience and perspective.
Re: (Score:3)
As a developer you do not necessarily have control over things like infrastructure and dependencies. The software solution might have dependencies in modules you have nothing to do with. You might not even be able to test the entire solution on your local system. OS dependencies might not be your decision to make. Cost, performance and/or time to market might outweigh your desire for a pure and easy to deploy solution. This is the real world.
The point is: In larger/complex de
Re: (Score:3, Insightful)
Why is someone who needs Docker a shitty developer?
Because if you can dismiss all new things that you don't understand as being only relevant to "shitty developers" you can attempt to justify your incompetence to your manager and keep your lowlevel job.
Manager: "Why aren't we using docker for our deployment environments?"
ArchieBunker: "Oh that's just a thing shitty developers do, don't look at it." :crosses fingers:
Re: Sounds like you're the shitty developer! (Score:2)
"Because if you can dismiss all new things that you don't understand as being only relevant to "shitty developers""
Then we wouldn't be on Web 11.0 "Now with Auto-Play Video Ads"?
Re: (Score:2)
I'm a hardware guy. I don't code but I can tell the difference between good and bad engineering practices. Can no one compile statically anymore?
Re: (Score:3)
I'm a hardware guy. I don't code but I can tell the difference between good and bad engineering practices. Can no one compile statically anymore?
A bunch of containers memory-sharing common components can use less resource than the equivalent statically linked processes. And modularity brings dependability and savings in lead times. Also deploying swarms of containers can flatten out load and optimise hardware usage.
Re: (Score:3)
There are developers who have written more than 3 languages you arrogant muppet. I personally haven't worked on a project that used less than 3 languages in over 2 decades and I have had code run for years on a daily basis without issue.
Re: (Score:2)
Having a use for Docker doesn't mean you're a shitty developer, but it certainly does enable those. In science, people noticed that nobody could get anyone else's shit code to work, so now the big thing is distributing containerized versions and calling it "reproducibility."
Re: (Score:2)
Docker is useful to give you some degree of portability between cloud providers, but hardly anyone uses it for that.
Re: (Score:2)
It's a preconfigured lightweight virtualized environment. When you're doing cloud deployments that's a pretty useful thing.
Whatever the fate of Docker might be, couldn't this tech be useful to all those hospitals that are being wiped out by ransomware because the controller for each piece of major medical hardware has to run on the Windows XP that was current when it went into FDA certification 25 years ago? Encase each of these instances in a VM bubble of its own, with only tightly specified communication paths to/from the outside, and you take each such environment out of the malware destruction zone.
Re: (Score:2)
Re: if someone could tell me what docker is or doe (Score:5, Insightful)
Developers who don't understand security get to bundle obsolete versions of libraries in their app images and you get to deal with the security breaches because they never get updated.
And occasion their front door inteacts with a bug in Docker so the attacker can get root on your system.
Now to be fair, there are dozens of responsible developers around the world using Docker who do care about professional responsibility and stay on top of CVE's in the libraries they use and update to be compatible with fixes. It's just that those folks do fine in a non-Docker environment as well.
Re: (Score:2)
Amen, brother.
No way a fire would melt all the backup drives in my closet, surely some of them would survive.
Re: (Score:3)
I still don't get it. Seems to me I just need to create a directory and modify the path variable to point there temporarily.
Re:if someone could tell me what docker is or does (Score:5, Insightful)
No, this affects C programmers too. Statically linking libraries is the devil, because one of those libraries may have a security issue someday, and if we dynamically link to them, why, we can fix that issue with one tiny update. Magic. Beautiful. In theory.
The reality is, of course, that dynamically linking everything creates a deployment nightmare, where you can never really depend on which of a bazillion possible variations your program might run into, and where a tiny, inadvertent breaking change might hose things. So instead, developers invented an even more convoluted process where instead of just statically linking libraries, the could now bundle up an entire filesystem, all dependencies included, along with the app. Woo, problem solved, at least for us!
It's seriously bonkers. Both sides have good points. Bundling everything up, either via static linking or using containers, does indeed mean that it's harder to deploy security fixes. But you also can't close your eyes and sing "lalalala" and pretend the deployment and compatibility nightmare doesn't exist. After all, a secure system is pointless if it also doesn't run the software you need.
Re: (Score:2)
Re: (Score:2)
Where do you find work in Ada?
The U.S. military?
Good ... (Score:5, Insightful)
Bankruptcy couldn't happen to a more useless ball of shit ...
Re:Good ... (Score:5, Funny)
Re: (Score:2)
I know. Just the other day I was thinking "I wish I could be completely at the mercy of my distribution maintainer to decide what version of software I get to run without hosing my system.
This whole Docker thing made Linux too easy. Real men fuck up their distro by trying to work around dependencies.
Re: (Score:2)
O RLY? purdue-pharma-accused-of-fueling-opioid-crisis-files-for-chapter-11 [npr.org]
Of course, it's a WIN for them, since it's just a sham to squirrel away their massive piles of money.
Lol, how? (Score:5, Interesting)
"Docker has already raised $272.9 million, but the company hasn't been profitable."
How do you piss away $272 million? What, exactly, cost that much?
I mean, I understand there are infrastructure costs, development expenses, and the crazy big budget for free coffee and snacks in the kitchen, but where did the other $270 million go?
Re: Lol, how? (Score:3)
Well, I get close with 500 employees, at $500k apiece. Remember, this is spread over six years. Plus rent. Infrastructure. Sales and marketing.
Re: (Score:2)
at $500k apiece
They don't cost that much, even after taxes and expenses.
Re: (Score:2)
In the Bay Area, at $150K starting grad salary, yeah easily.
Re: (Score:2)
So add 40% to 150k and you get.. 500k? Wait, how does that work?
Re: (Score:2)
Then it's bullshit all over.
First, it costs roughly 40% overhead to keep an employee in the building, cover their insurance, worker's comp insurance, HR overhead, etc.
Second, not all of those employees worked there that whole time.
Third, they are not all doing the same job, and even if they were, they wouldn't all be paid the same.
Re: (Score:2)
Re: (Score:3, Interesting)
"How do you piss away $272 million? What, exactly, cost that much?"
You ever tried to serious market any product? A single page ad in a magazine with a monthly run of about 50k subs, you're looking at around $5000+ for a single run. Now mulitply that and you're ad budget alone is close to $20m-$30 or more. We've not covered wages, average tech wage of about $100k for several hundred people ( plus excessive bonuses for middle management! ) and then there's infrastructure rental. I've worked with companies tha
Re: (Score:3)
People. Someone else pointed to Wikipedia: "Docker, Inc had over 423 employees in 2015" That was four years ago, and the company has existed since 2010. So the $272 million/423 math works out even if you assume started with much less headcount, haven't made much money, and so on.
Re: (Score:3)
It's way worse than that. Remember they have been burning through all their revenues as well.
When the housing bubble blew, the govt was able to shift all the rubbish loans into a 'bad bank' where they sat around at zero interest rates until their money printing had reinflated the asset bubbles. But I don't see how they do this with the corporate debt bubble. When the employees and customers have packed up and gone home, it doesn't matter how much money you inject into the carcass, the insolvent business is
Re: (Score:2)
"How do you piss away $272 million? What, exactly, cost that much?"
Petabytes and petabytes of AWS network egress from hub.docker.com. I've docker-pull'd quite a few TB over the last five years.
Re: (Score:2)
Imagine what a random community open source project could do with $270 million. Oh, no, wait, they'd just become Mozilla.
Poetry (Score:2)
"Lack of clarity about what?"
Is just a wonderful sentence.
Who knew you could make money by... (Score:3)
I mean, come on people !
Docker doesn't actually solve the problem it claims to solve anyway. Just ask anyone who's ever had libc compatibly issues.
Buyout time? (Score:2)
Sounds to me like Docker would make a perfect acquisition target for something like IBM or Google or whatever.
Just not Oracle. Hell even Microsoft seem to behave themselves these days. Anything but Oracle. Its just got to be a company thats prepared to leave it out in the FOSS world.
Docker Doesn't Make Sense (Score:2, Interesting)
Good riddance (Score:2)
The more I learn about Docker, the more I'm convinced it's a horrible idea.
It's the wrong, lazy way to approach what can be considered a real frustration, due to the Right Ways being "too hard".
Docker kind of defeats many of the core reasons for having (reusable) libraries in the first place. As well as coherent systems you manage and understand, and can update. It's not much of an improvement from a black-box binary blob and all the legitimate demonizing those get. But apparently if you make a bloated blob
Re: (Score:2)
i mean, a negative number divided by 10 is still a negative number...
Re:Wikipedia helps put part of it in context (Score:5, Interesting)
Worker costs are probably not what's keeping the business from being profitable. If this were true, then they've got a flawed business model perhaps. They also exist in 7 locations; 423 people won't even fit in the office space they have in SF.
Re: (Score:2)
Worker costs are probably not what's keeping the business from being profitable. If this were true, then they've got a flawed business model perhaps. They also exist in 7 locations; 423 people won't even fit in the office space they have in SF.
Maybe I'm being unimaginative, but I cannot imagine what 423 people are doing with Docker. How many kernel devs are there?
Re: (Score:2)
Maybe I'm being unimaginative, but I cannot imagine what 423 people are doing with Docker. How many kernel devs are there?
Why not? Can you do everything yourself? You'll have coders, managers, marketing, web design, internal infrastructure, HR, strategy, event organisers, training, enterprise customer support, legal, and C-suites at the very least and I'm being unimaginative as I'm sure there are far more departments than that.
There's no fixed list of kernel developers, but it would appear from a 2017 report that 15,600 individuals have contributed to the Linux kernel over a 10 year period.
Just for fun though... I cloned the g
Re: (Score:2)
You forgot testers. You probably want two testers per dev.
And layoff 90% of those employees, probably (Score:2)
Fornthe company to be profitable, they may need to burn a lot less money, which means getting a lot smaller.
Being open source and wildly popular, the Docker software could be just fine as a community-developed project with two or three people at Docker Inc providing leadership. Tens of thousands of companies use it, including Google, Amazon etc. Whatever needs to be fixed or added, somebody who wants it will do it.
Get costs down to below revenue by employing only enough people to manage the things that b
Re: (Score:2)
It doesn't have to be just a couple of people organizing an open source project. Docker could probably do quite well putting a nice face on Kubernetes and offering great support. Google ate their lunch in terms of development, but Google's big weakness is support.
Re: (Score:2)
Given the fact that NIMBYs are deliberately doing their utmost--quite successfully--to keep the costs very high in San Francisco, it may be time for companies to simply relocate to cheaper locales.
Well, that and I can't imagine how the Docker project could possibly require 423 employees. You can't really throw out a useful number without doing an actual analysis of the workforce, but just from a glance that number just seems high. Really high.
Re: (Score:2)
Managers are evaluated by how many people they manage. So the key is to convince someone to give you a bunch of money to hire as many people as possible, so you can be placed as highly as possible in your next job.
What those people actually do is beside the point.
Re: (Score:2)
Re: (Score:2)
Totally wrong. They are unprofitable because others can do what they have done (since it is based almost entirely on OS functionality) and others can also do what they were planning to do to monetize it, because it's not that complicated.
If you want to make money selling OSS then you have to be more competent than the next guy, and also better at marketing. Docker is neither.
Re: (Score:2)
This is something that is really misunderstood in the software industry. An idea can be "neat, usefully, and widely accepted" but still not marketable. Docker is the perfect example. There is nothing high tech there. Its basically nuts and bolts type stuff (and truthfully its not even that good) as you say all the heavy technical lifting is provided by the platforms.
The simple fact is the shops that really need orchestration features of Docker are already using a free (as in beer) alternative Kubernetes