##Securing your containers... Beyond Default

Docklock lets you encrypt your images and critical data, and allows a zero knowledge registery in third party registery services like quay.io. or docker public registery.

###Inspiration With recent emergence of heartbleed & poodle bug, we need to consider take extra measures to protect our assets. We liked the spideroak's zero knowledge idea to tackle on the very issue.

###Use case for end-user

• Protect your system, if critical applications like browsers get compromised, by providing container isolation.
• Protect your critical applications like bitcoin wallet, browser data and other meta data, if your system gets compromised physically.
• Takes a little compute/memory overhead.

####Usage

docklock encrypt <imageid> 
docklock decrypt <imageid> 
docker push <imageName> #Default workflow
docker pull <imageName>

###Workflow

Base image 
		-- Dockerfiles -- custom images -- containers
						  					-- data only containers (volumes) 
							 				-- normal epipheral containers (only for compute) 

A docker container can be seperated into three contexts:

• Dockerfiles
  • if application is small, all of images can be represented in this form. So just have to create password protected archive
• Image
  • Base Image, taken from popular registery
  • Properitetry Image, which takes this base registery, and builds a dev environment around it. (To be encrypted)
• Persistent Data
  • Basically volumes, in which your application stores data. (To be encrypted)
  • should be stored somewhere safe, like S3/Dropbox/AnyOtherStorageService with zero knowledge principle.
  • Or can stay in server... during runtime, which can be protected by firewalling the VM.

Working with gpg

• We use python gnupg(gnu pgp) library to encrypt the /var/lib/docker/aufs/diffs of particular image contexts.
• Docklock.py also contains, a flexible way to set up key in your preferred storage location.
• Each key is fingerprinted by email id of user.