As we have seen, trying to use the HOSTS file for anything but the smallest of networks would simply be too inefficient to function on a large scale. If all of the systems on the Internet had to synchronize their HOSTS files, then all available network bandwidth would be consumed by this process, leaving no bandwidth for other services like SMTP mail.
Actually, in the very early days of the Internet this mechanism was tried. There used to be a single HOSTS file available for downloading. Updates to the central HOSTS file were submitted via e-mail by network managers, consolidated together, and then the latest version of the file could be downloaded to each of the separate sites. This effort was abandoned after the advent of the Domain Name System.
One of the problems with the old method was that it relied upon a centralized system that would consolidate and distributed snap-shots of the Internet. In order to function, three separate actions had to occur whenever a change was made: first the local manager had to upload the changes, and then these changes had to be incorporated into the master system, and then the master database had to be retrieved by anyone who wanted to access one of the new systems.
Rather than have a central site manage the data on behalf of all the organizations on the Internet, DNS allows remote systems to use lightweight database queries to the local data at its source. By using a distributed naming service and query protocol, any DNS client can query any DNS server for hostname and address information.
For more information on the architecture of DNS, refer to section C.3.1 DNS' Hierarchy. For more information on the data that is stored in DNS, refer to section C.3.2 DNS Resource Records.
Copyright © 1997, Unoverica Corporation. All rights reserved. Unauthorized use prohibited.
Send comments to docs@unoverica.com