The networked radar air picture is built using input from radars in many different locations. In an ideal world, each radar can track every target continuously. However, the laws of physics do not permit this. Target fades, terrain blockage, and spurious signals all conspire to make the situation not ideal. As a result, in general, no one radar is able to create a complete surveillance picture - hence the need for networking. Various alternatives have been considered for radar networking. This paper attempts to organize and categorize the alternatives and quantify the differences between them. Three fundamental radar networking characteristics are considered: the data distribution process (robustness and capacity), data grouping approach (measurement-to-track and track-to-track), and data-sharing approach used.