NAT Traversal in Peer-to-Peer Architecture

Peer-to-peer networks are well known for file sharing between multiple computers. They establish virtual tunnels between computers to transfer data, but NATs makes it harder. A NAT, Network Address Translation, is a process which transforms private IP addresses, such as 192.168.2.1, into public addresses, such as 203.0.113.40. The idea is that multiple private addresses can hide behind a single public address and thus virtually enlarge the number of allocable public IP addresses. When an application in the local network establishes a connection to Internet, the packet passes through the NAT which adjusts the IP header and maps an external port to the computer which sent the request. When packets are received from the Internet by the NAT, they are forwarded to the internal host which is mapped to the port on which the packet was received, or dropped if no mapping exists. In this paper, we will introduce you to NAT and P2P, we will discuss the numerous ways NATs use to translate private IP addresses into public ones, we will discuss known techniques used to fix the problem and we will also present how popular peer-to-peer programs bypass NATs. This paper is written so anybody with a reasonable knowledge of networking would grasp the essentials. It is important to keep in mind that the traversal methods presented in this document work for UDP and TCP and require no manual configuration of the Network Address Translator itself.