A Research On Two-Source Network Coding

Network coding (NC) was firstly introduced at 2000 has the ability to increase the throughput and reach the maximum flow bound by allowing the intermediate nodes in the network to combine (mix) the arrived packets. Network coding has a big affect on the way that networks are designed and connected, it becomes an interested hot spot that many researchers highlighting and focusing on.In single-source network coding, the main goal is to find and characterize the maximum rate in which information can be transmitted from the source node to all sink nodes simultaneously. In multi-source network coding, characterizing the capacity region and information rate region is the main target of the researchers; this problem is not completely solved. Multi-source network coding is very difficult and complicated, it not easy to generalized, for this reason researchers use a special case of multi-source network coding, that is, two-source network coding.In this thesis two methods are suggested, the first method can achieve the capacity region and information rate region for two-source network coding problem. The suggested method depends on decomposing the multicasting operation to three steps and assign rates at each step separately according to its own multicast equirement. The first and third steps are multi-commodity flow problem; while the second step is the network coding problem. This method controls the sending rate for each source node. After applying the suggested method to many network topologies; we found from simulations results that the capacity region is achieved in addition to achieving the information rate region for two-source two-sink networks. Furthermore, the suggested method can deliver more throughput than ordinary network coding because it has routing step which exploits the small amounts of links capacities. In addition, increasing the links capacities will increase the throughput.The second suggested method in this thesis is the minimizing method that decreases the number of encoding nodes in the network. The suggested method is general, it can be applied for single-source and multi-source networks because it deals with each source-sink pair separately. This method has the ability to decrease the number of encoding nodes by selecting alternative path(s) to transmit information via it. From the simulation results, we found that this method can decrease the number of encoding nodes in the network; this method does not depend on the packets number or number of terminals in the network.