The major difference between the various P2P applications comes down to how the database of IP addresses is generated and administered. This dynamic influences the level of the hierarchy and the relative looseness of the P2P network.
In addition to using a different port from most other Internet-oriented applications, P2P programs have flown under the radar of many network administrators because they do not have many -- if any -- host addresses hard-wired into them. Where most networks used to guard certain ports but leave the others open, the smart network administrator now uses firewall rules to block all but essential ports.
The rigidity of the P2P structure is directly correlated to how centralised its database structures is. When the P2P program is first installed, it must have a way of contacting at least one other node in the network. From this node, the newly installed program will learn the addresses of others within the immediate network, progressively learning the locations of other peers.