5 ms·
I've generally thought Alex is the most pragmatic of the Container Kids. Flo seems to have a pretty decent business, something that is realistically lacking in
by neom 10y ago
I've generally thought Alex is the most pragmatic of the Container Kids. Flo seems to have a pretty decent business, something that is realistically lacking in the go to market at Docker, recently because they seem very emo. However, Alex/CoreOS has good technology that I believe is often overlooked. HashiCorp is to my mind always the dark horse, I still firmly believe containerisation is stop gap and part of the reason I campaigned hard internally at DigitalOcean to not deeply productize a container solution. Fully expect ten thousand down votes for this post. :)
- gtirloni 10y agoWho's "Flo"? Are you talking about HashiCorp's Nomad? Why is containerization a stop gap? Maybe answering these questions will "contain" the downvotes ;)
- neom 10y agoFAIR! Florian Leibert is the CEO of Mesosphere, I'm talking about the HashiCorp toolchain generally as an ecosystem, containerization is stop gap because I believe in event driven architecture/applications, and there are/will be new interesting and unique ways to implement that///cloud, however, I'm often wrong. :) :) I will say, I know all of those folks relatively well and they are all really really really good humans who I believe in and love.
- soulnothing 10y agoI was in the opposite position I was working at a hosting company in 2010. At that time I was heavily pushing containers, at the times Jails, Zones, LXC, etc. Using a full hypervisor stack for VMWare was not the best use of resources. I honestly haven't kept up with containers too much, other than Kubernetes, Misos, etc. I'm also thinking they're a stop gap. In the CI work flow it definitley helps with dynamic language, or slim executables. Where dependencies are seperate. But often I see infrastructure almost meshing with the code. A block, in the overall program design. I'm not sure the next step either, often as you said you're looking at a event or service driven architecture. It comes from the pain points you're trying to solve, and how you actually get your code running.
- kornish 10y agoWhat do you see potential successors to containers being? What are your thoughts on NixOS, which uses binary hashes to immutably declare dependencies?
- neom 10y agoNixOS is awesome and I lost a very long internal battle years ago to support it heavily. Very very pro NixOS. I've not been in cloud for almost a year now so I've not been following, however, my company builds fully towards AWS like this: lambda+kinesis+sqs+rds+s3+kms
- sjellis 10y agoThat makes sense to me. Right now, containerizing Web applications effectively adds a lot of complexity to the infrastructure, which means that hosting providers need to have a number of services in production, and the customers have to be willing to deal with that complexity. Lots won't, or can't. We use a mix of providers for different things, and I actually love DigitalOcean because of your focus on making it super-easy to run basic Linux servers at reasonable cost.
- jacques_chester 10y ago> I still firmly believe containerisation is stop gap It's a primitive that can be composed into other things. It's a better primitive for many cases than VMs are. If you're fretting about containers/not-containers, you're below the value line. What matters is application code that provides value. If you're thinking about how it gets launched and managed, you're losing time to someone who doesn't have to. Disclosure: I work for Pivotal, we have a fairly close interest in this topic through PivotalCF.
- neom 10y agoYou're still talking about VMs?! :) containers are stop gap, VM is old school, different things.
- jacques_chester 10y agoIn the sense that the uses of both have been about isolating apps from each other. The early fuss about VMs was about increased density -- but only because hardware used to be the boundary of isolation, meaning lots of idle capacity. Containers are a cheaper way to isolate, but the point was and still is isolation. Separate servers -> VMs -> containers. Same song, different instruments.
- neom 10y agoGood history lesson, thanks... you might want to look into why google created kubernetes if those are your opinions, anyway: et l'avenir?
- jacques_chester 10y agoI'm not sure what you're driving at. Orchestration is an example of taking a lower-level primitive and building on it. There are no real novelties in software architecture, just occasional shifts in algo-economics that cause previously burdensome ideas to become cheap and attractive. So predicting the future is premature at this point. It always is, but particularly at the moment. Everyone, including Pivotal, is busy trying to invent it.