How Docker broke in half

https://www.infoworld.com/article/3632142/how-docker-broke-in-half.html

by pauljonas • 5 years ago

302 194 5 years ago

NortySpock• 5 years ago

I really enjoyed reading this history of the creation of Docker and the company that supports it (Docker, Inc., previously dotCloud, Inc.)

randombits0• 5 years ago

I almost enjoyed it. I need a better ad blocker. That site sucks on mobile.

mikestew• 5 years ago

I have a better ad blocker, so now it's just white rectangles chasing me down the page, with a big white banner at the top that also chases page scrolls. Were it not for reader view, I wouldn't even have bothered with reading it. I'm dead serious when I say that I think I'll start using InfoWorld as my pi-hole test bed rather cafemom.com (which used to be my egregiously obnoxious go-to for ad-laden sites).

WolfeReader• 5 years ago

Using Firefox/Fennec with uBlock Origin works just fine for me! https://f-droid.org/packages/org.mozilla.fennec_fdroid/ https://play.google.com/store/apps/details?id=org.mozilla.fi...

threatofrain2• 5 years ago

Do you have reader mode, which is available on many browsers?

sorokod• 5 years ago

Or use Firefox Focus.

quantumwoke• 5 years ago

Good read, and touches on a lot of the pain points from that era. As someone who was lurking on HN at the time during the containerization boom, I think that the key failing of dotCloud/Docker was not capitalising on Docker Swarm almost immediately. Docker Swarm was touted almost from the start but the repeated delays gave it a reputation of smoke and mirrors left people scrambling for solutions.

I also clearly remember the multiple high profile spats that 'shykes had on HN which burned a lot of bridges. At the time he had a reputation for answering lots of questions on HN which helped a lot with community building. After those bridges were burned there was no one to speak for Docker as the developer mindsets shifted slowly towards Kubernetes. IIRC the Github PRs were also a source of contention as dotCloud corporatised.

To be fair, Kubernetes was a real slog to understand at the start and had a lot of competition; it was definitely not the same level of simple, direct technical solution that Docker was.

Interesting trip down memory lane and what a pivotal technology! Regardless of the rest Docker is a true cultural phenomenon and a testament to the insight of the creators working outside of the myopia of big tech.

bcantrill3• 5 years ago

Yes. Anyone who wants to relive an early ignition of one bridge should read the HN thread on the Rocket announcement.[0] No one is sympathetic with respect to the Rocket/Docker spat; both companies behaved poorly -- and it was increasingly clear that the energy at Docker, Inc. was going to spent fighting others rather than building a business.

Speaking personally, the fate of Docker, Inc. was clear to me when they took their $40M Series C round in 2014. I had met with Solomon in April 2014 (after their $15M Series B) and tried to tell him what I had learned at Joyent: that raising a ton of money without having a concrete and repeatable business would almost inevitably lead to poor decision making. I could see that I was being too abstract, so I distilled it -- and I more or less begged him to not take any more money. (Sadly, Silicon Valley's superlative, must-watch "Sand Hill Shuffle" episode[1] would not air until 2015, or I would have pointed him to it.) When they took the $40M round -- which was an absolutely outrageous amount of capital to take into a company that didn't even have a conception of what they would possibly sell -- the future was clear to me. I wasn't at all surprised when the SVP washouts from the likes of VMware and IBM landed at Docker -- though still somehow disappointed when they behaved predictably, accelerating the demise of the company. May Docker, Inc. at least become a business school case study to warn future generations of an avoidable fate!

[0] https://news.ycombinator.com/item?id=8682525

[1] https://www.hbo.com/silicon-valley/season-02/1-sand-hill-shu...

truffdog• 5 years ago

I remember having a ton of sympathy for CoreOS, Docker really was hard to integrate with as someone doing sysadmin automation stuff.

Docker went to war with the standard Linux init system and the kernel developers who maintained cgroups, their core system primitive. I don't really understand how that was supposed to work out.

danudey3• 5 years ago

I worked at A Thinking Ape (another YC alumni), and for the longest time we had people beating down our door to invest, but the founders always told us why they weren't going to take it: we didn't need money. We had money to run the business, but we didn't have a handle on what we would do with $10m or $100m. What do we need to do to get more users that stick around and generate profit? More and better advertising? Sponsoring YouTube content creators? Hiring 50 more developers? Acquiring other companies?

Without a clear idea of what specifically you need to spend money on, and how specifically you are going to spend that money to generate growth and value, taking money is a bad idea. And if the answer to that question is "so that we can keep paying our developers", then you're probably already doomed.

wpietri3• 5 years ago

raising a ton of money without having a concrete and repeatable business would almost inevitably lead to poor decision making

This for me is one of those "I feel like I'm taking crazy pills" lessons. It seems so obvious to me, and I hesitate to even estimate the vast amount of cash wasted in the last decade like this. But it's a lesson that an awful lot of people seem to insist on learning for themselves. Or not!

cdrini2• 5 years ago

Very well written article! Personally, I'm still waiting for docker swarm. The thing is enterprise doesn't care about developer experience. They don't need to; they pay developers so much, they have over qualified developers who can grok Kubernetes, and can hire more at a moment's notice. They care much more about whether it meets their goals. Smaller companies/orgs are what care about DX. They have likely very few, potentially not super highly qualified developers, and very limited resources. To them, something that their developers can set up, understand, and maintain, that's reliable and resilient, is a HUGE win.

clipradiowallet• 5 years ago

Personally, I'm still waiting for docker swarm.

What are you waiting for, is there something new in the works wrt swarm? I was a swarm fan early on(and secretly still am), but otherwise k8s is where I spend all my time at work.

cdrini2• 5 years ago

It never quite felt "ready"; it's development and docs always felt like an after thought. I guess it's more to do with marketing than the actual service. And I'm always afraid they're going to axe the thing. It probably is ready enough to give it a whirl.

zaphar2• 5 years ago

I think you vastly overestimate the skill level of enterprise developers. Enterprise may pay a lot but they also hamstring developers in some interesting ways which tends to attract people who are better at the political game than the tech. The high pay attracts skill just the wrong skill in many cases.

threatofrain2• 5 years ago

The truth is, Docker had the chance to work closely with the Kubernetes team at Google in 2014 and potentially own the entire container ecosystem in the process. “We could have had Kubernetes be a first-class Docker project under the Docker banner on GitHub. In hindsight that was a major blunder given Swarm was so late to market,” Stinemates said.
Craig McLuckie, Kubernetes cofounder and now vice president at VMware, says he offered to donate Kubernetes to Docker, but the two sides couldn’t come to an agreement. “There was a mutual element of hubris there, from them that we didn’t understand developer experience, but the reciprocal feeling was these young upstarts really don’t understand distributed systems management,” he told InfoWorld.

The article criticizes Docker Swarm as myopic, but IMO, there were only two possibilities for Docker to move forward; either they acquired Kubernetes, which was a possibility in this telling of events, or they won with their own Docker Swarm.

heavenlyblue• 5 years ago

“Didn’t understand developer experience”

I am wondering if anyone ever actually properly used docker swarm in any proper way? Especially as a small-scale developer.

doctor_eval2• 5 years ago

Yeah we used it in production, I’m not there any more but as far as I know it’s still running.

Swarm was good, easy to set up and run, but it had a lot of networking bugs. More than once we had to cycle the docker daemon on all machines which basically resulted in a rolling outage.

We’re trialling nomad at our new co, I was put off by the opaqueness and complexity of K8s, but apparently that’s just me.

brown9-2• 5 years ago

a great piece of trivia related to this point is that the original k8s developers tried to convince Docker to adopt the pod concept:

https://github.com/moby/moby/issues/8781

zzyzxd2• 5 years ago

“The biggest mistake was to miss Kubernetes. We were in that collective thought bubble where internally we thought Kubernetes was way too complicated and Swarm would be much more successful,” Jérôme Petazzoni, one of Docker’s first and longest serving employees, said. “It was our collective failure to not realize that.”

They were not wrong on saying that Kubernetes was very complicated, at least in some sense. In the beginning no one wanted to use it because they could easily setup Docker Swarm with minimal effort. This argument still pops up frequently on HN when there's a new post about Kubernetes.

I guess the problem was they didn't realize why Kubernetes needs to be that complicated. And if a system is complicated for good reasons, that's actually good business opportunity and tons of people and companies will be willing to make the effort to fill that gap.

gizdan3• 5 years ago

On this note, HN users and several HC employees (including some enterprise architects) I've spoken to love to mention Nomad and how it's much lighter and simpler to run, completely oblivious that many with bigger names have tried and failed to go up against K8s.

Kubernetes has become a speeding bullet, it won't be easy to catch and it has to run through a lot of walls to slow down for a competitor to do any damage to it's dominance.

schmichael3• 5 years ago

completely oblivious that many with bigger names have tried and failed to go up against K8s

Ha, I can at least assure you we're not oblivious. Nomad has a product manager from Mesosphere and most of HashiCorp's products integrate deeply with Kubernetes. We're well aware Kubernetes is the juggernaut to which every other scheduler has succumbed.

I believe there's room for both Nomad and Kubernetes. Whether as competitors or complements, having more than one vibrant project in the orchestration space will hopefully make all the projects better for users. Any one project has to make tradeoffs that improve the user experience for some while degrading it for others. For example Kubernetes has an IP-per-Pod network abstraction which provides an excellent out of the box PaaS-like experience for microservices in the cloud. On the other hand Nomad's more flexible network model more easily supports networks outside of flat cloud topologies whether it's small home raspberrypi clusters or globally distributed edge compute.

lenkite• 5 years ago

We will see if the next hot scaling/clustering/orchestration solution replaces k8s in a decade from now. Fingers crossed.

There have been lots of speeding, hard-to-catch bullets in the past.

scaryclam• 5 years ago

Thing is, Nomad isn't trying to go up against K8S. It's part of the Hashicorp family of products, so if you use a couple of those, it's easy to add Nomad into the mix, or just use K8S. While other competitors have tried to out-do K8S, Hashicorp seem to try to work with the broader eco-system, which means some will just use Nomad to keep it simple, but don't need to for Hashicorp to continue working for them in other areas.

throwaway20371• 5 years ago

The ironic thing is that Docker became successful only because of the tremendous hype they generated about themselves, and then failed to see the K8s hype train barreling towards them.

Swarm is still what 99% of companies need when they adopt Kubernetes, and is probably 1/10th to 1/20th the cost. Somebody needs to re-invent swarm as an independent project so that we can stop investing in K8s. The cost of K8s isn't even in the complexity, it's in all the millions of side-effects of the complexity, like security, fragmentation, and specialization.

cpuguy83• 5 years ago

Swarm is (and always has been) an independent project: github.com/docker/swarmkit

waspight2• 5 years ago

What is so hard about k8s? I think for example aws with fargate and all the networking there is far more complicated… i have never used swarm though.

angelzen• 5 years ago

k8s is just a bunch of processes working together. The downside is that if something goes wrong, it is unclear which process is faulty and how to fix it. A misconfigured DNS may take down the whole cluster, and the symptoms are that every process fails because the network is out. Difficult to backtrack to the fault source, there are a lot of configs pertaining to the network that could be faulty.

On the plus side, if someone else configured the cluster correctly for you (e.g. a Cloud service like GKE), then it's a breeze to use.

geocar• 5 years ago

If you're talking about a hosted k8s like EKS or a toy/single-node k8s, in 2021, then nothing: k8s is much better.

But if you're on-prem, and have a tonne of metal and just arrived in 2014 via a time machine, swarm was so much simpler and simpler that if you were a sysadmin who already had their own scripts -- their own jenkins-powered IC and git-hooks that built and deployed -- whatever it is you were building -- then swarm looked like a nice gradual extension of that, and k8s looked more like starting over and admitting defeat.

zzyzxd2• 5 years ago

Exactly.

Personally, k8s is one of the easiest system I learned to use in the past few years. It has many components, too many that one can't practically understand all of them well. But every component follows the same design principle, so there's a clear path for a noob like me to get the full picture step by step.

The first thing I read on k8s was controller pattern[0]. Then everything becomes so easy to learn. Something wrong with a pod? Find its controller and start troubleshooting from there. Oh, the pod is controlled by a replicaset? Check replicaset controller. The replicaset is managed by a deployment? You know where to go.

That's why I said it is only complicated in some sense. I saw a lot of people started learning this system by going through a list of popular components and their concepts. I would've gotten lost easily and probably given up if I did the same in the beginning.

[0] https://kubernetes.io/docs/concepts/architecture/controller/

piyh• 5 years ago

As a guy who's been running stuff with docker-compose and is learning k8s, the learning curve for k8s is a cliff. Every single layer of the stack gets added complexity.

aeyes• 5 years ago

Look at the history: In the beginning Kubernetes didn't have deployments - only replicasets, it didn't have ingress, it lacked the cloud provider integration you have today...

It was barebones like Mesos but they managed to focus on the right things early on. Once these features and a bit of an ecosystem started to build, the question if we should stay on Mesos with a bunch of homegrown management tooling or move to Kubernetes wasn't even a decision anymore. There was still a significant amount of code which we later just deleted because it became a first class feature or the community implemented something much better.

rualca2• 5 years ago

I guess the problem was they didn't realize why Kubernetes needs to be that complicated.

I guess this strongly depends on one's needs. I mean, if all you want to do is glue together a couple of nodes and have a few stacks running on them, do we really need all the bells and whistles?

I mean, with Docker Swarm anyone can get ingress right from the start on any bare ones cloud provider. Deploy your Docker Swarm cluster, slap a service like Traefik, and you're done. It's a 10min job. What else does Kubernetes brings to the table?

sgk2844• 5 years ago

K8S is great if you have Google problems, but most people don't and I think much of hype around it is much like the hype that existed a decade ago around "web scale" and "big data" and "NoSQL for everything". Docker Swarm is (with some use-case specific exceptions) more than sufficient for any shop running a few hundred nodes with a few thousand containers – and it is two orders of magnitude less complex than k8s.

You can read Docker Swarm's docs end-to-end in < 30 minutes, and have a cluster running in roughly the amount of time it takes to install Docker on three boxes.

Docker's failing was a failure to market Docker Swarm. Most people don't even know it exists, yet every single machine with Docker has it already! The last time I checked, they only had a single dev that was part-time allocated to supporting it.

So it's not so much that Docker failed to jump on the k8s bandwagon, so much as they built a wonderful thing that they never marketed and semi-abandoned, to focus on Docker Hub and other core Docker things they thought they'd be able to more directly monetize.

Fortunately, Docker Swarm remains a wonderful piece of software, even without particularly active investment, because it is a fairly complete offering for most use cases. And it seems to be growing in adoption.

bovermyer2• 5 years ago

I was dismayed when Docker effectively (and quietly) declared that they were no longer supporting Docker Swarm.

Docker Swarm remains far simpler to set up, run, and troubleshoot than Kubernetes.

I just can't choose it over K8s, though, knowing that it's unsupported.

MrOwen4• 5 years ago

Why only Google problems? Why not homelab problems where I'm managing upwards of 20 apps and I don't want to come up with custom automation to do things like cert renewal, tls termiantion, dynamic storage, and disaster recovery? What if I don't want to come up with my own system for organizing and grouping resources? Lots of homelab people (I assume) chain together hacky scripts and probably many things are not repeatable. With k8s, all my configs are in a single repo and I can spin back up my entire cluster on new hardware anywhere provided I've backed up my storage. How does Swarm handle all of these problems in a more unified or complete way that k8s does?

ritchiea• 5 years ago

One of the tough lessons of my years as a developer is many developers love to adopt Google scale solutions and Google practices because they’re from Google and they come with prestige and a feeling of proximity to the most visible technology company. Whether their problems are a good fit for Google scale solutions or not. It’s inevitable and it takes a lot of soft skills to argue against even if it’s dead obvious you’re going to be running a few instances of an application maximum. The simple solution is far less sexy than the Google solution.

manojlds• 5 years ago

K8S is great if you have Google problems

Borg is great if you have Google problems. K8s is not Borg.

danudey3• 5 years ago

The dream is to be able to take a software and yeet it into the cloud with minimal effort. That's not kubernetes at all.

That said, if your tooling or products or software supports kubernetes, which a lot more are these days, then it becomes a lot easier. A lot of companies these days are saying "Here's how to install it, or just grab this helm chart"; Gitlab, for example. We even had one piece of software we were testing which only had two options: cloud-hosted, or self-hosted on k8s. Nothing else was supported.

k8s doesn't seem to be any easier to configure, with the possible exception of microk8s (which has its own issues), but the amount of things you can do with it once it's up and running is compelling.

ljm2• 5 years ago

Unfortunately for Docker, I don't trust or want to use Swarm because of their attempts to monetize and because of the frequency that they push breaking bugs or breaking changes into production releases. To the extent that they made downgrading a paid option, not upgrading. And I'm in the process of switching everything I've got over to Podman now.

Swarm might be great, but the ecosystem around k8s is massive. So my two choices are: ignore Docker entirely and do it like I've always done (Heroku, chef, puppet, etc. just managing my servers myself), or buy into containers and go with K8S.

To that extent, I only ever deal with K8S at work, and for my personal stuff I don't go near Docker at all. Why would I when I can get an 8GB/4CPU VPS for $9/mo and serve everything I care about from that? I've got fail2ban, systemd, caddy, and plenty of resources.

I can host mysql on the same machine because there isn't a god-given law about hosting your DB on a separate box. Same with anything else.

No need for Docker or Swarm or K8s.

orthoxerox• 5 years ago

The problem with solutions bigger than `docker-compose up -d` and smaller than k8s is that they have a growth ceiling, real or imaginary. Sooner or later you will hit some limitation and will either thrash against it or have to upgrade to k8s. As a decision maker in an IT org, you can either:

- pick up Swarm or Nomad now, and then potentially be blamed for the expensive and costly migration to k8s - pick up k8s now, waste the time of your (voiceless) devs and ops on an overengineered solution, but retain access to the larger pool of technologies and a smooth upgrade path should the need arise

ghaff2• 5 years ago

Pretty much. Fundamentally Kubernetes “won.” Try to pitch your company on something else and you’re asking them to bet on something else “without particularly active investment” to use your words. As well as much less of a complementary software ecosystem.

dirkg3• 5 years ago

IMO the complexity of k8s tends to be overstated. Of course its a vast and complex system, but its also incredibly well documented and based on simple design principles. The reason people have such a hard time with it is -

- deploying k8s on bare metal is not easy. It was never really designed for this and to be fair Google has a vested interest in making you use GCP. EKS is nothing more than cobbled together scripts and you still need to do a lot of manual work

- k8s introduces a ton of terminology and new concepts (pods/service/deployment/ingress) not to mention the whole host of addons built on top like Istio.

It shouldn't take any decent dev more than a couple days to install minikube/k3s or use a cloud provider, play around and get comfortable with k8s.

zaat• 5 years ago

K8S based on distros like k3s or microk8s is a great and simple solution for manufacturing and managing edge devices, IoT solutions and similar products that are setup once on a single box that runs half a dozen or more things and are shipped to business clients.

It is very simple to spin a minimal distro based cluster. It is relatively simple to have it up and running your solution. Sure getting Kubernetes from the source and tailoring all the required components to run cluster with full HA etc' is a complex stuff, but building a private cloud based on vanilla Kubernetes isn't the only or the most common use case.

ajmurmann• 5 years ago

I suspect that there is a tendency to pick more powerful and complex tools and infrastructure than needed. For us as engineers it's easy to foresee some of the scaling-related, technical problems of the far future. It's so tempting to prepare for those, especially if you need to learn a new technology anyways and it looks great on your resume. In addition everyone knows intellectually that their own startup is unlikely to reach Google scale, but we do badly want to believed it will. So maybe the right bet is to always put your money on the more powerful and complex technology as long as the complexity is inherent and not incidental. Of course this is also limited to infrastructure. Consumer products and other markets would be totally different

technicolorwhat3• 5 years ago

I think docker itself had plenty of time to innovate. They could have solved the problem long before mainstream got a hold of kubernetes. They just didn't solve the actual problems users where having at that time, which was container orchestration across multiple servers. They went very very deep with a lot of drivers and API changes and what not. If I am not mistaken docker-compose wasn't even a part of docker first before it got bundled, someone else was solving docker's same-server orchestration problems.

Kubernetes is for a lot of organisations pretty difficult and way too much for a start. We used rancher for solving our problems which did a fine job, rancher got a worse when they made the move to Kubernetes.

In the beginning they (docker) also removed a lot of options like cpu throttling and configuration and such that LXC had from the start. They also had their own version (a shim) of pid 1, some point at the time. And other things that made it a little painful to properly containerise. I was often very frustrated by docker and fell back on lxc.

There was also something with the management of docker images like pruning and cleaning and other of much needed functionality that just didn't get included.

Also something with the container registries which I a this point can't remember (maybe deleting images or authentication out of the box or something that made it hard to host yourself).

Anyway I think it's failed because it failed to listen to its users and act upon them. They really had a lot of chances I think. I think they just made some wrong business decisions. I always felt they had a strong technical CTO that really was deep on the product but not on the whole pipeline from dev->x->prod workflows.

V99• 5 years ago

We used rancher for solving our problems which did a fine job, rancher got a worse when they made the move to Kubernetes.

(Early Rancher employee)

We liked our Cattle orchestration and ease of use of 1.x as much as the next person. Hell, I still like a lot of it better.

But just as this article talks about with Swarm, embracing K8s was absolutely the right move. We were the smallest of at least 4 major choices.

Picking the right horse early enough and making K8s easier to use led us to a successful exit and continued relevance (now in SUSE) instead of a slow painful spiral to irrelevance like Swarm, Mesos, and others and eventual fire-sale.

erulabs• 5 years ago

It's an incredible litmus test for developer v sysadmin. Developers think the complexity is unnecessary and so don't want to learn k8s. For sysadmins it's mostly just nice automation around all the chores we used to do with home-brewed bash scripts.

ffdf3r3rf2• 5 years ago

As a developer, I hate k8s, but mostly because my management does not understand difference between developers and devops. They changed our job titles to devops engineers and fired sys admins. I understand k8s is complex for people who specialize in it, try imagining as a developer, you need to setup private vlan ingress. At least, on regular Linux, I know how to setup ip tables or look up 1000s of tutorial.

K8s and devops has destroyed us developers' productivity, confidence, and morale.

goodpoint3• 5 years ago

I've never met a sysadmin that would prefer it over the bash scripts.

brazzledazzle2• 5 years ago

I wonder if some of the derision is at least due in part to k8s effectively forcing you to think about certain application characteristics upfront instead of addressing them later when you outgrow the single instance performance or suffer an outage.

dilyevsky• 5 years ago

What I haven’t seen mentioned here is at least by my recollection swarm was incredibly bug ridden especially in networking layer around ‘17/18 when k8s really started pulling away

makeitdouble• 5 years ago

I’ve seen a number of companies going for k8s and not other solution not because they needed the complexity, but because it was first party supported.

To me what killed Swarm is Google needing to spread k8s.

If you are eyeing at GCE but decide you want a bit more flexibility, going for a Google backed solution gives you k8s, and Swarm isn’t even on the choice list.

Choosing a third party means you’re on your own to bootstrap/make it live, or contract a vendor that might not live longer than you. That’s fine for companies who strategically choose to invest for the long term, sadly that’s not most of the companies out there.

wpietri3• 5 years ago

I guess the problem was they didn't realize why Kubernetes needs to be that complicated.

Does it?

It seems like there are three kinds of complexity involved here: the essential complexity of the total problem set, the essential complexity of a specific company's problems, and the amount of complexity experienced by a user as they try to solve those problems.

Let's stipulate that Kubernetes needs to be that complicated for Google's needs. I'm not totally sure that's true, but let's pretend for now. So we'll say you're right for the first kind.

But I'm still not seeing it for the other two.

For modest deployment efforts, Kubernetes is a giant pain in the ass. It's just way more complicated than needed for many non-Google use cases. I tried setting up Kubernetes for a bunch of personal stuff I run and it was egregious overkill. I persisted until my home lighting system went dark because some internal Kubernetes keys expired and it decided that just not working was a good failure mode. Last time I mentioned this on HN several people said, well duh, obviously you shouldn't run it just for yourself. And that's fine; every technology has an expected operational envelope. But there are an awful lot of projects out there that have between 0.0 and 0.5 sysadmins. For them it's too complex. Kubernetes is just too complex for many people's needs.

And the other side of that is complexity of experience. Modern processors are fantastically complicated. I'm on a mailing list for people in finance who micro-optimize everything to maximize response speed. Those people know incredible amounts about exactly what different processors do. And they have to. But most working developers know about 1% of that because a lot of people work very hard to hide the complexity from them. I don't think that really happened with Kubernetes.

At this point, it is what it is. It's much easier to scale a technology up than down, so I expect it'll always be like this. But For whatever comes after Kubernetes, I'd like to see much more attention paid to more common use cases.

coding123• 5 years ago

It's really unfortunate. I am a kubernetes user today, but I LOVED swarm. K8S' main differentiator is that specific intentional "cross-cloud" features such as ingress and volume mapping that help the implementing cloud understand the deployment topology. It's basically the new PaaS but instead of working on one cloud it can work on any.

I don't know that if Swarm had those features if it would really catch on like Kubernetes did. They didn't have the same budget as Google on that side of things.

SergeAx3• 5 years ago

I beleive that wide k8s adoption is the same dead end as JS frameworks. It is not actually needed and is too complex, so it introduced it's own class of pain. VMs were a great abstraction with very little added complexity, so was Docker. K8s doesn't abstract anything away compared with Docker swarm, just adding unnecessary complexity.

awinter-py• 5 years ago

ugh docker only ever did one thing right, allowed me to run postgres + mysql on my laptop without the DBs shitting on the rest of the laptop

if they had just released a chroot database runner, would have been as good as what they did

it has never been good at build or deployment. it has definitely never been good at cloud. even kube is still awful at most things

these systems have never been sure if they're configuration languages, buildsystems, plugin hosts, or operating systems, and so they've been bad at all 4

morpheos1372• 5 years ago

If software development was a real engineering discipline docker wouldn't exist. The idea that you need isolated environments to run/develop programs that need different versions of libraries is regressive. Way back in the 80s and 90s Microsoft had backwards compatiblity down. Now in the 2010s and 2020s every couple months we can expect a breaking change in the webdev, linux, python ecosystem. Try running a 3 year old python app outside a container. This foot gunning is dumb. But it sure does make work for software developers. Who go on HN and crow about how much value they create while messing around in docker or whatever.

rualca2• 5 years ago

The idea that you need isolated environments to run/develop programs that need different versions of libraries is regressive.

Containerization's value proposition isn't really the isolation part. That's nice and all, but it's not their main selling point.

The main selling point of containers is the fact that they solve the problems of packaging, deploying applications, and configuration, and they do so in a perfectly auditable and observable way.

morpheos1372• 5 years ago

Again if backwards compatibility were respected then packaging and deployment wouldn't be so complicated.

ghshephard3• 5 years ago

It's important to distinguish between the different flavors of containerization - Docker, and container orchestration - K8S. Docker is intriguing as a nice lightweight way of separating your processes - and is (mostly?) interesting to developers. K8S changed the entire trajectory of application management at scale.

We have about 4300 containers in production where I work (at this moment - it can change from hour to hour), and I don't believe "isolated environments that need different versions of libraries" was ever a driver, or even strong incentive for using containers.

Rather - the ability to rapidly add resources, resize resources, re-allocate storage performance dynamically, and use whichever cloud provider made sense (we've worked with AWS, Google Cloud, and now Azure) - and be able to migrate your environments between them in a period of a week - was what made containerization attractive.

Being able to type:

    kubectl scale deploy worker-background --replicas=300

And have 300 worker pods come online in < 60 seconds really changes the power of an individual. Being able to have that done all dynamically (up and down) without any human intervention - that's a game changer in some of the ETL heavy workloads.

If you've ever worked at a large company - think of the effort required to get 300 new 6 CPU/32 GB VMs allocated to a project you are working on. Now consider a company where pretty much all of the engineering / ops employees can do that from the bash prompt.

I'm just looking forward to the day where we can scale the performance of our storage from 500 IOPS/sec to 8K IOPS/sec dynamically as required (still requires a manual step with Azure - albeit a pretty straightforward one in their portal).

jiggawatts3• 5 years ago

This is a little bit hilarious to me, because I've been through this exact same "reinventing the wheel" cycle about 3 times now.

The cycle goes like this:

    "A mainframe takes too long to buy, it's too expensive!"
    "Here, have an LPAR, you can spin one up in seconds!"
    "Wooaa there! You can't just do whatever you want, fill in this paperwork to get an LPAR!"
    "Getting an LPAR approved takes too much paperwork!"
    "Here, have an Intel server instead! It's not a 'proper' machine, but you can have it in a week!"
    "Wooaa there! You can't just do whatever you want, fill in this paperwork to get a new server provisioned properly!"
    "Getting a server provisioning approved takes too much paperwork!"
    "Here, have a VMware virtual server instead! It's not a 'proper' machine, but you can have it in a day!"
    "Wooaa there! You can't just do whatever you want, fill in this paperwork to get a new VM provisioned properly!"
    "Getting the VM provisioning approved takes too much paperwork!"
    "Here, have a Cloud server instead! It's not a 'proper' machine, but you can have it in minutes!"
    "Wooaa there! You can't just do whatever you want, fill in this paperwork to get a new cloud IaaS VM provisioned properly!"
    "Getting cloud IaaS VM provisioning approved takes too much paperwork!"
    "Here, have a Kubernetes Namespace instead! It's not a 'proper' machine, but you can have it in minutes!" <-- you are here.

ghaff2• 5 years ago

If there were no need for isolated environments in Windows, VMware wouldn’t exist.

pjlegato• 5 years ago

Broken dependencies in 90s-era Microsoft systems was so common that it has its own Wikipedia article: https://en.wikipedia.org/wiki/DLL_Hell

goodpoint3• 5 years ago

Spot on, and this is seriously harming Linux distributions and ultimately making FOSS unusable for non-developers.

boucher• 5 years ago

I am a huge fan of Docker the tool, and during the prime of Docker the company they were doing some really impressive stuff. Swarm was dramatically easier to use than Kubernetes. The people working on Docker both in and outside of the company were top notch and they should be proud of what they accomplished.

It's too bad they didn't make the business side of things work, though I agree that at the time there was a certain feeling of too much money chasing an uncertain future for an open source project. I hope this next iteration of the company works out for them.

mountainriver2• 5 years ago

Swarm may have been easier to get going but from what I heard it had a lot of operational problems and lacked a complete feature set

debarshri2• 5 years ago

But did kubernetes really won the orchestration war or is it still an ongoing thing?

aynyc3• 5 years ago

It won. Right now it's a speeding train. To compete with it, you better come up with something that's mind blowing, or a culturally shift in development.

debarshri2• 5 years ago

I am not sure. almost 80% of the customer we talk to have troubles with kubernetes. They might be too small to operate it or something.

mountainriver2• 5 years ago

It has for the most part, nomad is it’s only real competitor now

endisneigh• 5 years ago

I'm curious - does anyone have a story on how they used Docker Swarm for something at scale (>100 nodes), but later migrated to K8s successfully at the same scale?

technicolorwhat3• 5 years ago

There is no mention on LXC/jails which is ofcourse the biggest inspiration for docker, if I am not mistaken it also used lxc under the hood. That was already a good'ish product and but hard to configure and not for the mainstream at the time. Docker introduced AUFS and downloading of images which was added. I always saw docker as a properly marketed nicely ribboned lxc but mediocre implemented since they removed a lot of options at that time that were super useful like cpu limiting etc from std lxc.

stuff4ben• 5 years ago

Wow this was a great read! I remember taking a kubernetes class in late 2015 and the instructor then was saying that Kubernetes had already won over Docker Swarm and Mesos.

If I were at the helm of Docker today, I'd focus on Enterprise customers. Provide an on-prem version of DockerHub and you'll convince thousands of companies tired of Artifactory to switch immediately. Don't get me wrong Artifactory is pretty nice, but it's a bear to run at scale on an enterprise level. DockerHub already is handling that traffic, so there's no reason to think they can't port that to an on-prem offering.

schipplock• 5 years ago

Nexus3 can be a docker registry. It’s built-in.

merb2• 5 years ago

why even bother? there is quay, which is way better than dockerhub or the open source integration. and even than there is also harbor, nexus3 and as you mentioned artifactory. ALL of them can be bought with support from way bigger companies than docker inc. so why even bother?! Docker Inc. basically failed by trying to do everything on their own.

IneffablePigeon• 5 years ago

We used Tutum before it was acquired by Docker and turned into Docker Cloud. Whilst it probably would never have scaled to our complexity needs indefinitely as we grew, it was very clear that as soon as it was purchased the product stopped going in a positive direction.

The only changes that they really rolled out were replacing the older, perfectly functional UI with something superficially shinier but much, much harder to use.

Not only was it just hard to see what you wanted to see once they'd halved the information density and got rid of any visual hierarchy, you had absolutely crazy things like a slider for the number of instances of a service that would instantly apply with no confirmation and, even crazier, responded to the scroll wheel. We once doubled our number of production pods and didn't notice for a few hours, because someone scrolled up the page and their cursor went over the slider. Lucky they weren't scrolling the other way.

Anyway, I lost a lot of trust in Docker as an organisation that knows what people value about their products as a part of that.

sbmthakur• 5 years ago

The combination of huge amounts of venture funding, a quickly growing competitive landscape, and the looming shadow of cloud industry giants all wanting a piece of the pie created a pressure cooker environment for the young company to operate within.

I actually got an AWS ad on the same page while reading this comment. It's definitely hard to compete when your competitors(with lot of capital) can easily integrate your tech into theirs in no time.

swlkr• 5 years ago

I wonder if there's more room for things like fly dot io but with things like minio/s3 or lambda included as well for bandwidth savings, kind of an "aws lite"

jokethrowaway2• 5 years ago

The real problem is that docker swarm is nowhere close to being production ready and it's full of gotchas and design decisions that make little sense as an end user. They simply couldn't make a product good enough for the enterprise, like k8s did. I have similar reservations towards docker; sure it was a mostly novel concept (I saw live when Solomon announced docker and relatively few people knew at the time about Linux native containers or about bsd jails) but it so full of quirks and gotchas and bad APIs that it's not smooth sailing.

Unfortunately they couldn't make docker into a product (or they would have faced the wrath of the OSS community) so they probably felt stripped for time on both docker and on swarm, which made for an unpolished experience.

I've been using swarm for a small project for funz and because I didn't want to run the full k8s (even k0s or k3s would have been heavy for my use case) and because I had bad experiences with nomad.

My list of complaints: - Stacks are not working 1-1 with Compose - docker machine is a separate binary - You can't pull the latest of a docker image or the cache hit will just fail to redeploy your changes (you're forced to do a sort of blue-green deployment or just update a tag everytime) - You have to configure, run, garbage collect a registry from scratch - docker-flow-proxy should be included out of the box - I need a way to integrate secrets

Don't get me wrong, there's an amazing amount of progress in docker swarm (and I'm running it in production, for free, on a 5$ machine). Years ago it was even worse but it's not something I would ever recommend to my employer. I hope we'll get there eventually.

rejectedandsad2• 5 years ago

Hykes does acknowledge that there were tensions between the Docker and Google teams at the time. “There was a moment when egos prevailed. A lot of smart and experienced people at Google were blindsided by the complete outsiders at Docker,” Hykes said. “We didn’t work at Google, we didn’t go to Stanford, we didn’t have a PhD in computer science.

I found this passage interesting. Are these tensions common?

Fiahil2• 5 years ago

Yes, I graduated from the same school as Salomon a few years later, and I got a lot of “you’re just a random dude from nowhere, go back to the playground” vibe during my time in California.

It’s quite an American attitude to judge your entire life and skills solely based on the college you went to. I didn’t encounter such attitude with my fellow European and Australian colleagues.

Nowadays, I don’t care that much because I’m not the little, defenseless, graduate student I was, and I do enjoy bringing a Google-Standford-Californian down to earth when they are bragging about something they have no idea about.

rejectedandsad2• 5 years ago

Depressing. I went to a basic state school and work at a normal company (not Google) and it feels like I’ll be destined to stay a failure forever.

zaphar2• 5 years ago

I never graduated from college. I also worked at Google for 7 years and never once got that vibe from my coworkers. It probably varied by office and I was in the Chicago office. But even on my somewhat frequent visits to New York or California offices I still didn't get that vibe.

boulos• 5 years ago

We didn’t work at Google, we didn’t go to Stanford, we didn’t have a PhD in computer science.

Amusingly, except the "work at Google"-part, Joe and Craig don't have those other qualifications either. Brendan did get a PhD and was even a professor, but I dunno, he gave up on that.

jchonphoenix• 5 years ago

I don't have the same recollection Solomon does and I was in the room. The Google guys had an opinionated view of the world based upon experience and wanted logical justification for why their method was not superior.

Solomon's "outsider" comment is ironic since multiple of us in the room from Docker did, in fact, formerly work at Google and go to Stanford. He walked into that room assuming Google would have those prejudices and that's all he could see.

greatgib• 5 years ago

I don't like so much the idea pushed by the article author that it is partly the fault of the software to be open source!

At least, the founders of Docker are honest and clear in my opinion:

<< Hykes disagrees with this assessment. “I think that is wrong and generally speaking the core open source product created massive growth which created the opportunity to monetize in the first place,” he said. “Lots of companies monetize Docker successfully, just not Docker. There was plenty to monetize, just Docker failed to execute on monetizing it.>>

There is still a thing that is missing there: the article let think that docker was an innovation out of nowhere and they had an unique idea that was kind of spoiled.

But, one has to remember the context of the period when Docker was created:

It was a time were cgroups and "namespaces" of network, process, ... were the hot new things inside the Linux and everyone was thinking about the concept of "container", one way or another.

There were already chroots and a lot of persons were already working on "app"/"module"/"package" systems using aufs/overlaysfs layers.

So, this was the moment and a few competitor were emerging like lxc and docker. Docker was very good to take the of the media cover and hype and so to become the dominant way to have "containers". But, at the beginning, there was nothing particular in the technology and lots of competitors could have done it if it was not Docker.

meatmanek• 5 years ago

Certainly my recollection is that containers didn't enter the zeitgeist until Docker came around. LXC and OpenVZ had existed for years before Docker, and Jails/Zones had existed in BSD/Solaris for years before that. Mentioning the long-standing existence of these tools was a favorite pastime of Docker detractors in the early days.

Really, Docker's innovation wasn't the ability to run processes in a container (as mentioned before, that technology had existed for years already); it was the Dockerfile and the Docker hub/container registry. Prior to Docker, the way you'd build a container was by running debootstrap or unpacking a tarball/zip file of a root image. Take a look at https://developer.ibm.com/tutorials/l-lxc-containers/ . As far as I recall, there was no LXC equivalent to `docker build` or the registry except maybe wiring up Packer or home-grown shell script and throwing a tarball in S3 or something.

xorcist2• 5 years ago

The Docker registry is what made Docker popular. (One could even suspect that free hosting for developers wins popularity contests.)

What made Docker really explode was when it ran on Mac and Windows. The cloud runs on Linux, but desktop computers generally don't, mostly for cultural reasons.

This made for an explosive combination. There are probably similarities to Github here.

indolering• 5 years ago

Yeah, Docker nailed the UX right but screwed up the implementation and community building.

Red Hat specifically “weren’t part of the community, they never rooted for the success of the Docker,” he said. “The mistake on our end was desperately wanting them to be part of the community. In retrospect we would never have benefited from that partnership.”

I understand how it seemed like other open source companies were dictating terms, but Docker's implementation was crap: one fat daemon running as root, no sandboxing, no cryptographic signatures, and a total lack of integration with the rest of Linux. Red Hat gets to define how containers should work on the back end because it pays for ~7% of all kernel development (compare that to Canonical's ~1%) [0]. But instead of cooperating with others in the Linux ecosystem, Docker only begrudgingly adopted interoperable standards while Red Hat, CoreOS, Intel, and others implemented fundamentally better solutions.

We didn’t work at Google, we didn’t go to Stanford, we didn’t have a PhD in computer science. Some people felt like it was theirs to do, so there was a battle of egos. ... Kubernetes was so early and one of dozens and we didn’t magically guess that it would dominate.

Google had experience running containers at scale and was using K8s as an opportunity to retool their internal proprietary solution. It sounds to me that Google engineers tried to explain to less experienced team the best path forward ... and Docker took it as an insult.

Sure, not everyone needs all the features of K8s, but Docker could have built a simpler interface on top of K8s and benefited from Google's investment. Instead they built a half-baked competitor that didn't work well. Whoops!

Docker delivered a slick developer UX but their architecture was antithetical to how Linux developers thought it should work. Then they spurned offers of help from both the hyperscalers and the dominate Linux enterprise vendor [1] and chose to build inferior solutions for those markets. Now they are stuck selling their pretty UI because that's the only thing Docker does fundamentally better than anyone else.

[0] https://www.linuxfoundation.org/wp-content/uploads/linux-ker...

[1] https://www.redhat.com/en/blog/red-hat-continues-lead-linux-....

mikesabbagh• 5 years ago

kubernetes makes simple deployments look complex, and complex deployments look simple. Docker would be my choice for a simple deployment (like server + db) Only use kubernetes when you need to do complex stuff, it does make things look simple then.

manishsharan• 5 years ago

I am surprised everyone is blaming K8s for eating Docker's lunch but from my experience in the banking industry.. it was cloudfoundry that stole the market away from Docker swarm.I don't think I ever saw a Docker swarm in production; IT departments standardized on cloudfoundry to manage their "private/on-prem cloud".

mixmastamyk3• 5 years ago

There's a lot of "dark" matter and energy in this industry that doesn't get much attention, even when it may be the majority of mass in a particular area.

jrochkind1• 5 years ago

How many employees does Docker has/had? I feel like if I got a couple hundred million dollars in funding, I could make it last many many years, producing what Docker has produced. But of course this feeling isn't based on any entrepeneurial experience, so I'm probably wrong. But that seems like an awful lot of money to shoot through.

AlexCoventry• 5 years ago

VC investors don't give you that money to sit on it. They want you to scale, and quickly.

ilaksh• 5 years ago

I think what happened with Docker is that it was just too good at what it was originally intended to do and too easy to use for that.

Programmers hate things that are easy to use (by the way I am a programmer, and no I don't feel that way) because if they admit they use them then they might be accused of being users. And of course no programmer will admit this. They don't realize it because it's actually a subconscious psychological issue.

And so what programmers started to do was immediately make it much more complicated and at the same time, take it completely for granted. It was so useful it was like a floor to walk on. And so people started giving it the same level of respect they give a floor.

So for those reasons, Docker became very uncool. But at the same time it was incredibly useful. Solution: make something just like Docker but not Docker, which hip people will be allowed to use without any shame. Make it a bit more complicated and only run on the cool expensive hardware.

jrm4• 5 years ago

As someone only casually knowledgable in this area (e.g. I use docker for some of my home automation thingies, I still haven't wrapped my head fully around what Kubernetes is beyond something like a way to manage a bunch of similar containers? VMs? eg)..

Docker, the kabillion dollar company, never had a chance. We all recognize the potential and actual problems when some very crucial piece of infrastructure is only free/open source and not maintained -- but this is what happens when it is "overmaintained" I.e. it makes sense to invest big in something that is good and could be crucial; but only if you can both keep up quality AND stave off competitors -- and in this case, competitors aren't just other companies, but free software in general. It's kind of funny, in a way, Docker made itself a target for "good competition" by being so visible, in a way that e.g. "curl" didn't.

Either way, I always knew "Docker the kabillion dollar company" was a stupid stupid bet. Small company, great bet. But this? Destined to fail.

jeffthechimp• 5 years ago

Happened to a pair of my Dockers once.

fulafel• 5 years ago

Anyone know any good metrics or analysis about trends of Kubernetes vs other ways to run containers in the not-huge-scale dev world which is 99% of software?

zxcvbn4038• 5 years ago

My issue with Docker Inc has been with the support. I started to give them money once but pulled back after it became clear the support experience was essentially “customer support thyself”. I don’t have to pay anyone to have my support ticket ignored, I can do that myself for free.

Anyway they are past the point of no return now. Their creations will undoubtedly outlive them - probably Queen Elizabeth also. Anything they pull back will be replaced by the community. Anything new will get aped by everyone else trying to make a name for themselves. Containers are more or less a generic commodity now.

korginator• 5 years ago

Use the right tools for the job. I continue to run small to medium scale deployments with a couple dozen containers with Swarm. It's smooth, and just works. Kubernetes would work too, and I use it for newer deployments.

Heck, I use plain old docker-compose to manage production servers with half a dozen containers, trouble-free for a few years now. Compose v2 seems promising, I've been using it for a while and it's pretty stable too.

Swarm could have been an industry in itself, a de-facto standard for orchestration if it had been marketed and supported right. It was a lot more stable than Kubernetes in the early days and things just worked with docker while you'd have to kludge a few things to get Kubernetes doing its thing right.

say_it_as_it_is• 5 years ago

The desire to change the world was stronger than the desire to make money while doing it

bacan• 5 years ago

Kubernetes is an absolute joke. Even after being set-up right, it rarely works as it is supposed too.

hnarn• 5 years ago

I have always been somewhat skeptical of Docker, for example due to how horribly it has always worked on Windows. My first real exposure to containers was when getting into Red Hat certifications and being introduced to Podman, and with the benefit of hindsight I don't really understand the benefit of Docker today when comparing the two. Podman seems just as easy, more restricted and more in line with what some might call "Unix philosophy".

Of course when it comes to distributed environments I'm sure Kubernetes still reigns supreme, but for relatively simple container setups I don't think there's a need to even run Docker anymore on Linux.

Obviously, I'm still not well versed in containers so this might all be overlooking something.

mixmastamyk3• 5 years ago

Yes, it's common for better designs to come along after insight has been gained by trial and error.

hda111• 5 years ago

Docker has with no doubt much better UX. It just works and it’s very easy to use. For podman is seems you need a bit more knowing what you are doing. You need to know iptables. With Docker your database container is reachable in the Internet automatically. Not possible to lock it down with iptables rules. (Could be a disadvantage too.)

How Docker broke in half

Related Stories