Yes, rkt App Container runtime is based on cgroups and namespaces. Same as runtimes of Docker, lmctfy, systemd-nspawn, lxc or (almost) any other container technology for Linux.
The big difference is on specification of images. rkt introduced the App Container specification for its images: https://github.com/appc/spec/blob/master/SPEC.md#image-archi...
[0] http://keepass.info/help/v2/dbsettings.html#compression