s3-signer is intended to be an aid in building secure cloud-based services with AWS. This library generates cryptographically secure URLs that expire at a user-defined interval. These URLs can be used to offload the process of uploading and downloading large files, freeing your webserver to focus on other things.
- Minimal depedencies
- Web framework agnostic
- Reduces web server load
- Simple API
- Ideal for AJAX direct-to-s3 upload scenarios
S3 Query String Request Authentication
AWS Specification
Signature = URL-Encode( Base64( HMAC-SHA1( YourSecretAccessKeyID,UTF-8-Encoding-Of( StringToSign ) ) ) );
Haskell Implementation
module Network.S3.Sign ( sign ) where
import Crypto.Hash.SHA1 (hash)
import Crypto.MAC.HMAC (hmac)
import qualified Data.ByteString.Base64 as B64
import Data.ByteString.UTF8 (ByteString)
import Network.HTTP.Types.URI (urlEncode)
-- | HMAC-SHA1 Encrypted Signature
sign :: ByteString -> ByteString -> ByteString
sign secretKey url = urlEncode True . B64.encode $ hmac hash 64 secretKey url
{-# LANGUAGE OverloadedStrings #-}
module Main where
import Network.S3
main :: IO ()
main = print =<< generateS3URL credentials request
where
credentials = S3Keys "<public-key-goes-here>" "<secret-key-goes-here>"
request = S3Request S3GET "application/zip" "bucket-name" "file-name.extension" 3 -- 3 secs until expired
S3URL {
signedRequest =
"https://bucket-name.s3.amazonaws.com/file-name.extension?AWSAccessKeyId=<public-key-goes-here>&Expires=1402346638&Signature=1XraY%2Bhp117I5CTKNKPc6%2BiihRA%3D"
}
-- Quick and dirty example
type FileID = ByteString
makeS3URL :: FileID -> IO S3URL
makeS3URL fileId = generateS3URL credentials request
where
credentials = S3Keys "<public-key-goes-here>" "<secret-key-goes-here>"
request = S3Request S3GET "application/zip" "bucket-name" (fileId <> ".zip") 3
downloadFile :: Handler App (AuthManager App) ()
downloadFile = method POST $ currentUserId >>= maybe the404 handleDownload
where handleDownload uid = do
Just fileId <- getParam "fileId"
-- Ensure file being requested belongs to user else 403...
S3URL url <- liftIO $ makeS3URL fileId
redirect' url 302
- Configure S3 Bucket CORS Policy settings
- CORS Docs
<?xml version="1.0" encoding="UTF-8"?>
<CORSConfiguration xmlns="http://s3.amazonaws.com/doc/2006-03-01/">
<CORSRule>
<AllowedOrigin>https://my-url-goes-here.com</AllowedOrigin>
<AllowedMethod>PUT</AllowedMethod>
<AllowedHeader>*</AllowedHeader>
</CORSRule>
</CORSConfiguration>
- Retrieve PUT Request URL via AJAX
type FileID = ByteString
makeS3URL :: FileID -> IO S3URL
makeS3URL fileId = generateS3URL credentials request
where
credentials = S3Keys "<public-key-goes-here>" "<secret-key-goes-here>"
request = S3Request S3PUT "application/zip" "bucket-name" (fileId <> ".zip") 3
getUploadURL :: Handler App (AuthManager App) ()
getUploadURL = method POST $ currentUserId >>= maybe the404 handleDownload
where handleDownload _ = do
Just fileId <- getParam "fileId"
writeJSON =<< Data.Text.Encoding.decodeUtf8 <$> liftIO (makeS3URL fileId)
- Embed FileReader blob data to request
- Send upload request
var xhr = new XMLHttpRequest();
xhr.open('PUT', url /* S3-URL generated from server */);
xhr.setRequestHeader('Content-Type', 'application/zip'); /* whatever http-content-type makes sense */
xhr.setRequestHeader('x-amz-acl', 'public-read');
/* upload completion check */
xhr.onreadystatechange = function(e) {
if (this.readyState === 4 && this.status === 200)
console.log('upload complete');
};
/* Amazon gives you progress information on AJAX Uploads */
xhr.upload.addEventListener("progress", function(evt) {
if (evt.lengthComputable) {
var v = (evt.loaded / evt.total) * 100,
val = Math.round(v) + '%',
console.log('Completed: ' + val);
}
}, false);
/* error handling */
xhr.upload.addEventListener("error", function(evt) {
console.log("There has been an error :(");
}, false);
/* Commence upload */
xhr.send(file); // file here is a blob from the file reader API
How to read file data from the browser
- Why do I keep getting 403 forbidden when I attempt to upload or download from a pre-signed URL?
- Ask yourself the following:
- Are my keys specified correctly?
- Did I configure the CORS settings on my bucket properly?
- Still trouble? Make an issue
- Ask yourself the following:
- Why are my URLs expiring faster than the specified time?
- Ask yourself the following:
- Is my server's clock synchronized with AWS? See wiki for NTP info
- Ask yourself the following:
- Why didn't you use HMAC-SHA256?
- It's 30% slower, and for all intents and purposes no more secure than HMAC-SHA1 (no known vulnerabilities exist for it, to my knowledge). Plain SHA1 is a different story. Collisions can be found, but there is no known way to apply those to HMAC-SHA1.
- For the curious SHA-1 is broken
- For the paranoid (Schneier quote from same article above)
- Relevant SO Post
This attack builds on previous attacks on SHA-0 and SHA-1, and is a major, major cryptanalytic result. It pretty much puts a bullet into SHA-1 as a hash function for digital signatures (although it doesn't affect applications such as HMAC where collisions aren't important).