Replicated documents for the world-wide web
Henning Pagnia, Oliver Theel, Iwik · 1997
The number of users of the World-Wide Web (WWW) increases on a daily basis, thereby consuming more and more of the Internet's restricted resources. A large fraction of the workload consists in retrieving WWW documents on remote servers. Since WWW documents are generally infrequently updated but extremely often retrieved, providing multiple copies of these documents is the almost natural way to cope with the resource problem. By installing copies nearby points of frequent use, precious bandwidth can be saved. Furthermore, the load associated with the retrieval of "hot-spot" documents can be shared among multiple servers. Additionally, document retrieval and document update can be offered with high availability due to an inherent tolerance against site and communication failures. However, the correctness of the multiple copies must be guaranteed. A broad spectrum of protocols which enforce a certain notion of correctness have been developed which address this problem. In this paper, we d...