The rapid uptake of the Internet has resulted in a far larger demand for machine IP addresses than originally catered for in the existing network protocol design. It is no longer possible for every machine that is connected to the Internet to possess its own fixed IP address. Accordingly it is now common for many networked machines to be provided with dynamic IP addresses – an address that frequently changes.
Node addressing is a particular concern within P2P systems. The use of dynamic IP addresses can cause problems as they can make machines difficult to reach. A node on a peer-to-peer network might possess a particular IP address one day, but another the next. Consequently, unless other nodes are informed of the new address, that node can become lost from the network.
A common strategy has been to maintain a database of the current IP addresses of the nodes within the network. This is usually used in combination with a semi-centralised architecture, where the central index node maintains the database. With a decentralised architecture this becomes more difficult as the database would have to be distributed over all the nodes of the network and maintaining consistency would become an issue.