[jdub]

I'm surprised and a bit disappointed that this new platform still exposes users to the idea of a container host. In Carina lingo, you have to create a "cluster" and choose the number of "segments" in it.

Joyent's Triton avoids this completely: Their whole data centre is a Docker "host", and you never have to care about it. The way it should be.

(I'm not paid by Joyent, nor am I a customer -- I just like what they've done to push the model forward.)

[jnoller]

This is honestly fantastic feedback and spot-on for the level of abstraction I want (and will) aim for. This is an early Beta, so things like this are top of mind.

[girvo]

That's excellent to hear, because I believe that Docker/containers will truly take off once developers can treat them as the "highest level" of computation; ie. no hosts needed, containers are all that are worked from!

[nzoschke]

Containers are here to stay and almost certainly will be the new abstraction of computation.

My big question is around a service like AWS Lambda. Is that not already the logical conclusion of container based computation? If magic units of computation can run instantly on demand, what more do you need?

No hosts, but also no OS images, and no specific containerization tools!

[IanCal]

It's a massive shame that PiCloud went down as that had containerized environments, extremely fast startups and auto scaling. It was one of the few things I've used that really solved my small-scale data processing problems simply and cheaply.

> If magic units of computation can run instantly on demand, what more do you need?

Some control over flow, scaling and batching wrap everything up for me. Startup times for my code are non-zero even if the environment is, and adding on queues with a "batch grab" means I can scale things far more sensibly (I can cram a lot of stuff into a single matrix mul if I can pull 100 items at a time from a queue).

I really, really miss picloud :(

[jnoller]

We rebuilt picloud as a foray into the getcarina.com space - https://github.com/cloudpipe/cloudpipe we're going to be bringing that back now that carina is landed.

[jdub]

Cool!

[DannoHung]

Your assessment is accurate, unfortunately Triton does not address the issue of data locality, which is a pretty big problem unless you're already building your storage backend to be Smart Data Center aware.

To that extent, I'm not sure that tossing out the idea of hosts is great yet.

[pbowyer]

> Triton does not address the issue of data locality, which is a pretty big problem unless you're already building your storage backend to be Smart Data Center aware.

Can you expand on this please? I was not aware of the problem.

[DannoHung]

To put it as simply as possible: Triton does not have magic block storage. If you want to write anything that persists past a container's lifetime, it is tied to the machine instance the container was running on unless you manually move it around.

It's not any worse than you'd have if you were starting from scratch with your own hardware, but Bryan Cantrill is firmly against magic block storage. His suggestion is that you rewrite your persistence layer to know about SDC or "deal with it".

It's not necessarily that bad since a lot of containers don't need state, but my personal axe is that I work with databases almost exclusively.

[nzoschke]

I'm not disappointed. It's great to see Rackspace competing in container services.

But I agree... At this level of abstraction ECS is far more attractive because of the rest of the AWS services and ecosystem.

Joyent deserves a lot of credit for removing the extra layer. However Triton is missing key functionality like a good load balancer.

Abstractions are important, but pure functionality is more important. How can anyone beat AWS?

(Disclaimer: I am building open source tooling in top of ECS, and have published a head to head review of ECS vs Triton)

[abrookewood]

Can you link to that review please? Triton sounds very interesting.

[nzoschke]

Sure! In summary Triton is really awesome but it needs a load balancing service for starters...

http://stackshare.io/convox/amazon-ecs-vs-joyent-triton

[abrookewood]

Excellent article - thanks. If you do a follow up, I'd love to see some performance comparison between the two approaches.