Using DHTs to Untangle the Web from DNS

The marriage between DNS and the Web, while initially fruitful, is now seen by many as a mutually unhealthy union. DNS’s original goal was practical and limited: allowing users to refer to machines with convenient mnemonics. As such, it has performed admirably. However, with the advent of the Web and the resulting commercial value of DNS names, profit has replaced practicality as the dominant force shaping DNS. Legal wrangling over domain ownership is commonplace, and the institutional framework governing the naming system (i.e., ICANN) is in disarray. Commercial realities have transformed DNS into a branding mechanism, a task for which it is ill-suited. A linked, distributed system such as the Web requires a Reference Resolution Service (RRS) to map from references, such as Web URLs, to actual locations, such as the IP address(es) where the referenced object is stored. In the current Web, references are URLs with a hostname/pathname structure and DNS serves as the RRS. The host-based nature of URLs—which ties references to specific locations and hard-codes the path component of the reference—makes content replication and movement hard. Thus, the Web would be better served with a new RRS, one that does not embed location into the reference itself. The experiences with DNS suggest three basic design principles for any such RRS, which we now articulate. Location-independent references: References should not contain information about the location of the content and should easily generalize to the case of widely replicated content. Along with many others, we argue that a reference, like any abstraction used for indirection, should remain constant as the referenced object moves. With current Web URLs, however, hyperlinks break when the referenced content migrates since the hostname and/or the directory path to the object changes, and no mechanism exists for updating referring hyperlinks. Contention-free references: There is contention over domain names because they are used as the basis for URLs. Social problems include “name squatting”, “typo squatting”, and lawsuits over trademark infringement [6]. Moreover, there are many examples of misleading DNS-based URLs, such as http://www.martinlutherking.org, a Web site that, contrary to expectation, is decidedly hostile to Martin Luther King, Jr. All of these tussles result from DNS names having become branding elements. Rather than attempt to solve the