Pail is an API for your files.
It's goal is to provide you with http access to a directory on your linux filesystem and any metadata you want to add.
You can GET, POST, PUT and DELETE with a simple curl command:
# POST a file
curl -i -F file=@funny_cat.gif http://pail.example.com/
# GET the file
curl http://pail.example.com/funny_cat.gif
# DELETE the file
curl -X DELETE http://pail.example.com/funny_cat.gif
#PUT a file
curl --upload-file another_cat.gif http://pail.example.com/another_cat.gif
Object storage stores data as objects with metadata, identified by a uuid. An API is employed to access the objects.
Pail is a REST API for a block storage filesystem. It has all the power of an object storage API without the need for the underlying object storage hardware.
WebDAV(Web Distributed Authoring and Versioning) is a unique protocol with many features for reading and writing documents.
Pail uses HTTP for all it's communication, providing maximum accessibility to any language or system. POST, PUT, GET, and DELETE are the only methods you need to know.
NFS(Network File System) is a distributed, open filesystem protocol that you would mount and treat like a local disk. It is extremely valuable if you need complete control over the files and their heirarchy.
Pail runs over HTTP, so there is no mounting or special ports to open. It works through firewalls, proxies, and NAT.
iSCSI(Internet Small Computer System Interface) is an IP protocol for sending SCSI commands over the network. You can do everything from mount a disk to boot from iSCSI. It is an extremely powerful storage bus.
Pail is very lightweight and easy to install. The only requirements to get started are Linux and Ruby.
Pail does not come with an encryption or authentication mechanism. If you plan on running vBucket over the Internet, its very simple to host behind Apache or Nginx. These web servers can provide basic authentication, SSL, reverse proxy.
At it's core, Pail is designed to serve up files. If you want to host files that are accessible via REST API, you can simply run vBucket as a service and point it at a directory on your linux system. You can upload and download files over HTTP with a multitude of languages, libraries and tools.
Pail can serve as a repo for linux packages. GET requests from yum or apt-get to the API are easy for vBucket.
You can also publish new packages to your repo by doing a POST or PUT. Pail can be configured with a post_upload script to run repo indexing commands after an upload has completed.
In addition to Linux repos, you can also serve gems from Pail. Simply use the post_upload in the configuration to perform a gem generate_index on your vBucket directory.
There is a lot of movement in this area right now and containers are all the rage. As container repos become more standardized, it should be possible to host containers much like RPMs in a yum server. Upload your container and any metadata, and rebuild the index with a post_upload script.
Hate storing your base Vagrant boxes in Dropbox? I do. Pail is a great way to keep them organized and accessible.
- File metadata - add metadata to your files!
- File query - list files based on metadata!