In this MSc Thesis, the principles of Network Coding will be reviewed, ranging from the description of the model to the typical applications. Also, two solutions related to privacy and security issues will be analyzed by taking as starting point an encoding node performing independently Network Coding in multiple flows. An attacker with traffic analysis capabilities could infer whether a set of incoming and outgoing data packets belong to a given flow simply by analyzing the linear dependency between them. Recently, Wang et al. have proposed a simple deterministic mechanism known as ALNCode to solve this issue by mixing (intersecting) the multiple flows in order to hide the correlation between the incoming and outgoing packets in each flow. In this Thesis, we characterize this solution by using combinatorial tools over projective spaces that deliver an exact expression for the general intersection probability in a given dimension. In addition, we extend the analysis using the properties derived from the intersection graphs defined by the generated subspaces. Finally, using error correcting coding, we analyze the special case of erasure data packets in the network and how this affects the performance of Network Coding.. The purpose of this project is the study of the performance of network coding modelled on undirected graphs, including the effect of different graph topologies. It is also meant to present the state of the art of those techniques both from a theoretical and an applied point of view