A Survey of Internet-Scale Namespaces
Rohit Khare, University of California, Irvine

Abstract:

The layered architecture of Internet applications is articulated through a slew of name/address boundaries. An opaque locator at one level becomes a structured locator for the one below. A popular web browser, for example, allows URLs to be typed into an "Address" field; but in turn that URL is decomposed and resolved against several more namespaces, such as scheme, host name, and path. This proceeds recursively down to Ethernet hardware IDs and port numbers; and all the way up to identifying human beings and vocabularies.

Inspired by the classic Eames film, "Powers of Ten", this talk begins at an airline reservation page and zooms in and out to introduce a host of Internet-scale namespaces embedded within an everyday application. Furthermore, we will describe several distributed resolution strategies, from Domain Name System to Address Resolution Protocol. Finally, we attempt to identify the cross-cutting issues to go beyond merely mechanical scaling to large numbers of names, to reach Internet-scale: architectures that work across organizations, across space, and across time.


TWIST 99

Irvine Research Unit in Software
Information and Computer Science
University of California, Irvine
Irvine, CA 92697-3425