how to make a peer to peer video conference system using webRTC APIS

1. Video Conference
Our goal is to make a Video Conference system in
our solution such that developers can communicate
with each other. Therefore, we first try the most
common solution which is using flash player at the
client side and streaming server at the server side.

2. OpenTok
One of the famous platforms that implement this
architecture is OpenTok which allows developers to
integrate live, face-to-face video directly into their
own website.
However, OpenTok is not free and very expensive
to use it inside a social network that may contain
thousands of users.

3. Streaming server vs peer to
peer
Therefore, we think about building our own system
but the problem was that streaming server is very
expensive too.
We try to found a peer-to-peer solution without a
streaming server so that each browser can send
and receive stream from other browser peer to
peer so we can save the cost of using expensive
streaming server.

4. WebRTC (Real Time
communication)
WebRTC has implemented three APIs
◦ 1. MediaStream API (getUserMedia)
◦ 2. RTCPeerConnection
◦ 3. RTCDataChanal

5. MediaStream API
(getUserMedia)
getUserMedia is available in Chrome, Opera and
Firefox Nightly/Aurora and can be used in Internet
Explorer using Chromo Frame plugin.
the getuserMedia API is responsible for access the
camera and microphone .presents synchronized
streams of media.
For example, a stream taken from camera and
microphone input has synchronized video and
audio tracks.

6. The getUserMedia() method takes three
parameters:
◦ 1. Constraints object.
◦ 2. A success callback which, if called, is passed a
LocalMediaStream.
◦ 3. A failure callback which, if called, is passed an
error object.

7. Here is some sample code

8. Signaling: Session Control, Network and
Media Information
WebRTC uses RTCPeerConnection to
communicate streaming data between browsers
(peers), but also needs a mechanism to coordinate
communication and to send control messages, a
process known as signaling.

10. 1-First choice is to use XHR and the Channel API
as the signaling mechanism (google app engine)
and this is the easy way and most common used
but it is not free and you have to pay to use this
service.
2-Second choice is to implement signaling
mechanism on your server using websockets

11. WebSockets using node.js
Node.js is a platform built on Chrome's JavaScript
runtime for easily building fast, scalable network
applications. Node.js uses an event-driven, non-
blocking I/O model that makes it lightweight and
efficient, perfect for data-intensive real-time
applications that run across distributed devices.
Node.js is used to run JavaScript without a browser
so you can use it to create a server side application
and create WebSockets.

12. Implemetation of WebSockets
java Node.js
pro - A multi-threaded language.
- It is an old language, have
a good community and
there are many resources
on the internet. - It is a
strong typed language with
can save a lot of error and
java good IDEs.
- It is built completely on
asynchronous callback
architecture when it receive a
request it execute it in a callback
function asynchronous on
another thread so we create
thread only when we execute
callback function not for every
connection like java.
con It is a multithread language
and every connection to a
WebSocket will need a
thread this will need a high
performance server.
- Node.js use JavaScript which is
not strong typed language which
can cause a lot of errors. - Very
difficult to debug.

13. Our decision
We decided to use node.js platform because it is
very efficient and doesn’t need a high performance
server so we can save the costs.

14. Scenario
Step 1: Each user connect to a socket on the
server using socket.io API.

15. Step 2: We need to add some data in sockets at
the server side so we can identify who is
connected on the server, who is online, and busy.
Socket is JSON object we add the following data to
it
1-Socket.devos.username=null;
2-Socket.devos.reciever=null;
3-Socket.devos.status=0;

17. Step 3:
1. If we want to send a message to a user you have
to add some header data which are the sender
user_id and the receiver user_id so at the server
read the message and search to the reciever socket
by user_id and forward the message to the receiving
user.
2. assume that browser 1 is Alice and browser 2 is
Bob .Alice want to make a video call with Bob so she
first send a message that contain her user_id as a
sender and Bob user_id as a receiver and type of
the message is call_request.
3. Then Bob response with a message contain his
user_id as a sender and Alice user_id as a receiver
and message type call_response and action is
accept

19. Step 4:
After Alice receive Bob response
1. she create a peer connection and create and offer and send
it to Bob as session description which contain information
about the video format and other data.
2. Bob receive the session description offer and create a peer
connection and add the remote session description to the peer
connection and then create and answer as session description
and send it back to Alice.
3. Bob use the ice server framework, which used in webRTC
peer connection API to get the shortest path to Alice and send
here his IP and port number as message called candidate
message.
4. Alice receive the answer session description and add it as a
remote session on her peer connection.
5. Alice use the ice server framework which used in webRTC
peer connection API to get the shortest path to Bob and send
here his IP and port number as message called candidate
message.
6. Peer connection complete setup after both of them receive
the candidate message the a stream flow peer to peer
between them.