Reference server implementation in Go of tus: the open protocol for resumable file uploads
Go to file
Felix Geisendörfer 9df1b91d68 Change license 2013-03-20 22:00:04 +01:00
scripts Bash script now correctly puts 2 parts. 2013-03-20 17:53:45 +01:00
src/cmd/tusd Comment on route change 2013-03-20 21:56:33 +01:00
.gitignore Ignore tus_data dir 2013-03-18 11:06:55 +01:00
README.md Change license 2013-03-20 22:00:04 +01:00

README.md

tusd

If content is king, you must make no mistakes acquiring it. tus provides the infrastructure for fast and reliable file uploads for your website or mobile app.

Sounds interesting? Get notified when it's ready: http://tus.io/

Motivation

It's 2013, and file uploading on the web is still an unsolved problem. There is a distinct lack of full stack open source software that allows developers to provide their users with the experience they deserve.

The tus mission is to make file uploading more reliable, faster and a better user experience. Instead of building yet another black box service, we are dedicated to providing an open source solution to this problem.

Roadmap

The initial goal for this project to come up with a good and simple solution for resumable file uploads over http.

  • Defining a good http API (first proposal created)
  • Implementing a minimal and robust server for it (in progress)
  • Creating an HTML5 client
  • Setting up an online demo
  • Integrating Amazon S3 for storage
  • Creating an iOS client

Future features will be based on your feedback. A few potential ideas:

  • Alternative transfer mechanisms: FTP, UDP, E-Mail, etc.
  • Security: Authentication Tokens, HTTPS, etc.
  • Support for running tusd instances in a geographically distributed cluster (reverse CDN)
  • Alternative storage backends: Cloud Files, Dropbox, etc.
  • More clients: Android, PhoneGap, etc.
  • Service integrations: Transloadit, Zencoder, Encoding.com, Youtube, Vimeo, Facebook, AWS Transcoder, etc.
  • File meta data analysis
  • Thumbnail generation

Once the project matures, we plan to offer a hosted service and support contracts. However, all code will continue to be released as open source, and you'll always be able to run your own deployments easily. There will be no bait and switch.

Getting started

Requirements:

Installing tusd:

Clone the git repository and cd into it.

git clone git@github.com:tus/tusd.git
cd tusd

Running tusd:

Run it with go:

go run src/cmd/tusd/*.go

HTTP API

Below is the proposed HTTP API for resumable file uploading.

POST /files

Used to create a resumable file upload. You may send parts or all of your file along with this request, but this is discouraged as you will not be able to resume the request if something goes wrong.

Request Example:

POST /files HTTP/1.1
Host: tus.example.com
Content-Length: 0
Content-Range: bytes */100
Content-Type: image/jpg
<empty body>

Response Example:

HTTP/1.1 201 Created
Location: http://tus.example.com/files/24e533e02ec3bc40c387f1a0e460e216
Content-Length: 0

The Location header returns the <fileUrl> to use for interacting with the file upload.

PUT <fileUrl>

Request Example:

PUT /files/24e533e02ec3bc40c387f1a0e460e216 HTTP/1.1
Host: tus.example.com
Content-Length: 100
Content-Range: bytes 0-99/100
<bytes 0-99>

Response Example:

HTTP/1.1 200 Ok

HEAD <fileUrl>

Request Example:

HEAD /files/24e533e02ec3bc40c387f1a0e460e216 HTTP/1.1
Host: tus.example.com

Response Example:

HTTP/1.1 200 Ok
Content-Length: 100
Content-Type: image/jpg
Range: bytes=0-20,40-99

The Range header holds a byte range that informs the client which parts of the file have been received so far. It is up to the client to choose appropiate PUT requests to complete the upload.

A completed upload will be indicated by a single range covering the entire file size (e.g. Range: bytes=0-99 for a 100 byte file).

GET <fileUrl>

Used to download an uploaded file.

Request:

GET /files/24e533e02ec3bc40c387f1a0e460e216 HTTP/1.1
Host: tus.example.com

Response:

HTTP/1.1 200 Ok
Content-Length: 100
Content-Type: image/jpg
[file data]

Prior art:

FAQ

Who is behind this?

Transloadit Ltd is funding the initial development. However, our goal is to build an active community around this project, so contributions and feedback are more than welcome!

Why not upload to Amazon S3 directly?

Amazon S3 has several limitations that we consider problematic:

  • The minimum chunk size for multipart uploads is 5 MB. This is by far too large for use under bad network conditions.
  • Throughput to S3 is often too slow for high bandwidth clients.
  • S3 is a proprietary service. Having an open, vendor agnostic API allows you to treat storage as an implementation detail.
  • The lack of uniform HTML5, iOS and Android clients that can be easily used to add reliable file uploading to any application.
  • While there is some support, S3 was not designed to be used in a browser environment.

S3 is an incredible offering, but we feel that it leaves much to be desired when it comes to offering the best file uploading experience to your users. We can build something much better.

License

We are still trying to figure out what license to use. MIT or Apache seems most likely at this point.