Project Proposal CS4220/6335 Motivation Two different approaches can be used to deal with the transmission error in the networks. One is Automatic Repeat Request (ARQ), and another one is <a href="/keyword/forward-error-correction/" >forward error correction</a> (FEC). Most of common protocols (HDLC, TCP/IP, etc) are using ARQ to ask for retransmission of the lost data packets. However, in the case of distributing real-time multimedia data, the ARQ mechanism will result in considerable delay and jitter which are not allowed in such applications. Our objective is to improve Quality of Service in real-time networks using the method of <a href="/keyword/forward-error-correction/" >forward error correction</a> (FEC). Due to sensitivity to delay, real-time communication is usually based on unreliable protocols such as UDP, since retransmission of lost packets takes too long. FEC provides a promising solution to the problem in that errors are corrected at the end point without the need to wait for the retransmission of a packet. Our attempt to solve this issue is to apply the technique of FEC and adjust the packet size and various such parameters to come up with reliable as well as efficient transmission of data. FEC The data to be transmitted can be divided into three priorities: Priority 1: The data that may not be lost during the transmission. Priority 2: The data can be lost, but the loss will definitely affect the quality of service. Priority 3: The data whose loss will slightly affect the quality of service. According to the requirement of data importance, three types of FEC are given as follows: Type 1 (for Priority 1): Use higher redundancy- for example, produce one redundant packet for each one transmitted, or one for every two, or two for every three. Type 2 (for Priority 2): Use lower level of redundancy- for example, produce one redundancy packet for every three, or one for every four, etc.. Type 3 (for Priority 3): Perform no FEC. All 3 types will be compared. There has already been some amount of work done [3] but the authors mention that refinement of FEC is required and that is what we are attempting to do. Experimental Setup Our goal is to first have a system with 1 Server and 1 Client which is transmitting data to and fro. Depending upon the bandwidth, network noise and network congestion we will try and implement a version of the FEC method to account for dropped and out of order packets so as to make it more efficient than retransmission. The system will be tested by transmitting a picture and comparing the efficiency of using FEC and a na ve UDP clientserver system. The system will ultimately be implemented on the emu-lab. Solution FEC in its various forms is to be studied. We shall be trying various combinations of transmission for e.g. If a picture can be sent using 10 packets then we could try sending 11 packets where 1 is used for error correction. Similarly we could send smaller packets but 20 of them and 1 more for error correction. Both these and similar approaches have to be experimented with in order to come up with an efficient solution. In one sentence an algorithm has to be developed and tested in order to mak...