Cryptloader

This is a proof of concept application that takes drag&dropped files, encrypts them on the client side and uploads crypted file to the server.

Encrypting is done using Standford Javascript Crypto Library and it actually is not the interesting part of this PoC.

The files are read using HTML5 File API, serialized using JS typed arrays and moved between server and client using Websockets.

Client side javascript is pure Vanilla-JS and the server is implemented using node.js. It requires just one module, Websocket-Node and it can be installed using the npm.

I have tested the server with nodejs version 0.8.14 and the client side with Firefox 16.

File serialization

Files are serialized using typed arrays and Int32Array ArrayBufferView. The file slices are read to the ArrayBuffer, which is then serialized to a string representing a javascript array of integers. This string is then encrypted and sent to the server.

Deserialization process creates an array of Int32Arrays and turns them to one big Blob object. The blob is then read to a data uri and the data uri is written to href attribute of a anchor tag.

As everyone should imagine, this is fairly big process and has a lot of overhead. So when you test it out, I suggest not to try it with big files...

Installation and running

1. Create directory files into the repository root
2. npm install websocket
3. node server.js
4. Point your browser to http://localhost:8888

Websocket protocol definition

All messages use the same data structure:

{"type": "message name", "data": "anything"}

"type" is always a string defining the reason of the message. "data" is message specific set of data. It can be a string or json object.

Uploading a file to server

Client can initialize the upload by using message type incomingfile where data is an object containing transfer id and number of slices in the transfer.

{"type": "incomingfile", "data": {"id":"randomly-generated-id","slices".3}}

Server responds with message type readyfortransfer with data containing the id and next number of file slice.

{"type": "readyfortransfer", "data": {"id":"randomly-generated-id","slice":0}}

Client can then send the requested slice using type fileslice with data containing transfer id, slice number and data.

{
    "type": "fileslice",
    "data": {"id": "randomly-generated-id", "slice": 0, "data": "foobar"}}

Server will response with either type readyfortransfer or transfercompleted if all slices are received. The transfercompleted data will contain the id and an filename which can be used to retrieve the file from server.

{
    "type": "transfercompleted",
    "data": {
        "filename": "name-on-server",
        "id": "randomly-generated-id"
    }
}

Using the filename from the transfercompleted message, client can now receive the file from the server by sending a message type get with data containing only the filename.

{"type":"get","data","name-on-server"}

Server will respond with message type incomingfile with data containing the id, which is the filename and number of slices in the file.

{"type":"incomingfile","data":{"id":"name-on-server","slices":3}}

Client can then request the slices with message type getslice with data containing id and slice number.

{"type": "getslice", "data": {"id": "name-on-server", "slice": 0}}

Server will send the requested part of the file with message type fileslice containing the id, slice number and file data.

{
    "type": "fileslice",
    "data": {"id": "name-on-server", "slice": 0, "data": "foobar"}
}

Client can continue to ask slices using the message type getslice, until all slices are received.

On error the server will send a message type error. The data will contain something about the error. It may be a string or an exception object.

{"type":"error", "data":"Something went wrong"}

License

The code is licensed under the GNU GPL version 3 or higher.