Domain names (panel session, abstract only): more questions than answers

The DARPA domain name project raises several interesting questions about the underlying assumptions and issues involved in providing a hierarchical naming scheme in an environment the size of the Internet. For example, the project was motivated by the growth of the Internet. However, because the proliferation of hosts took place within the sites belonging to the Internet (due in part to an increase in the number of workstations and other small machines), and not necessarily in the number of sites, a less general solution that maintained a centralized and fully distributed site table, with each site running a name server for its hosts, would have been sufficient. In addition, the distributed implementation of the name to address database uncovered several fundamental naming and addressing issues that had been ignore by the original centralized approach. For example, name servers and resolvers must be able to select a reasonable Internet address for hosts connected to the Internet at multiple points. The decentralized solution also raises questions about authority over host names.