Comparison chart for different continuous deployment tools for AWS

I recently needed to conduct a research on the various continuous deployment tools available.

My results are summed up in this table:

	Chef	Puppet	Ansible	CodeDeploy	Docker
What is it	A master-worksation service with OO DSL for deployment, uses git (ruby based) web UI on enterprise edition	A master-worksation service with OO DSL for deployment (ruby based) web UI on enterprise edition	Yaml based script, Push architecture, no clients (python based)	AWS deployment tool UI from AWS console	A container for a software environment so that the deployment env matches the dev env similar to AMIs
How to use	Install both client and server. Build the script trigger the workstations	Install both client and server .Build a manifest in the puappetmaster and then run the puppet script on every station	Install server. Create a playbook in yaml	Install on the servers and then use web UI to configure	Create a new (or use a perdefined one as basis) container. Install what you need. Commit it as a container. Upload to hub. Install it wherever you want
Learning curve	Need to learn the scripting language	Need to learn the scripting language	Need to learn the scripting language (but a simple one)	Very simple streightforward	Fast. Just a few docker commands
Price	Free with enterprise edition (6$ per node)	Free with enterprise edition (112$ per node)	Free	Free for now	Free open source
Scaling	Checf server deploys to multiple stations easily but no push	puppetmaster deploys to multiple stations easily using push	Ansible quick deployment to multiple severs	No prob	infinite
Environment support	All	All	All	All	Linux (based on linux containers)
On-premise	Can be deployed on on-premise servers	Can be deployed on on-premise servers	Can be deployed on on-premise servers	EC2 only	No prob
Supported langs	All	All	All	All	All
Cons	Takes some disk space (especcialy server). No push feature	Complexity of scripts. Steep learning curve. Expensive enterprise edition	Very simple - less control. No code reuse	Must use github (for now) Less environment configuration options	Need to build entire image. Slow deployment.
pros	downloadable playbooks, strong community, cheap (not really necessary) paid edition	Fast, reusable, widely used, big community with downloadable modules	Fast, simple, no clients, downloadable playbooks	Great EC2 integration. Automatic built in CD support. Very simple.	Community containers, full environment (very safe), simple (no need to learn a new language)