mreferre

[ Disclaimer: Author of the blog here ] I knew this blog was very polarizing so this is a fair comment. I have called out advantages and disadvantages in the first scenario ("The pros of this approach is that the customer is decoupling heavily from the infrastructure underneath thus enabling better portability of their entire platform. The cons of doing this is that the customer is signing up for a high amount of work.") and, in retrospect, I should have called them out explicitly for the last scenario (I have just assumed readers could infer they would be the opposite). I did mention towards the end that "customers [that can benefit from the third model ] are coming from an Amazon ECS background, they are all in into Amazon ECS and AWS" but I take to heart the suggestion that I should have been more explicit. Thanks.

Sorry, yes it has been down for a while. My basic hosting isn't sized to handle north of 8000 hits in a couple of hours :) (I did not expect this much as on average my site has a few hundreds hits per day).

@snockerton thanks for the kind comments. No that was NOT (by any means) how I have approached the post. Arguably the attempt of humor in the article and the picking of a misleading title may have misled the readers to think I was dismissive with the technology. But that wasn't the intent. I do find Lambda (and Serverless in general) very intriguing and I agree on the potentials.

Sorry that was a typo. I meant to say "how you'll do things post containers".

Corrected. Thanks.

Yes I agree. That makes sense. My point was more around the fact that all of the orchestrator tools out there aren't meant to be run on a "resource management framework" like Mesos.... they provide the whole end-to-end feature set. Marathon on the other hand has been engineered specifically to run on Mesos.

That's why I am defining "surgery" the idea of opening up the other tools to literally remove pieces of code to make them run on top of Mesos.