Implementing a Reliable Digital Object strongArchivestrong.pdfVIP

 Implementing a Reliable Digital Object Archive Brian Cooper, Arturo Crespo and Hector Garcia-Molina Department of Computer Science Stanford University f cooperb,crespo,hectorg @DB.Stanford.EDU Extended version ABSTRACT contains pointers to the different materializations of a given An Archival Repository reliably stores digital objects for long document (e.g., the postscript version, the plain text version). periods of time (decades or centuries). The archival nature The usual way of updating this set is to write a new pointer of the system requires new techniques for storing, indexing, into the set object, or to delete a pointer from the object. Be- and replicating digital objects. In this paper we discuss the cause the write-once policy forbids such changes, managing specialized indexing needs of a write-once archive. We also collections of objects using sets requires new storage struc- present a reliability algorithm for effectively replicating sets tures. Furthermore, these new structures require specifically of related objects. We describe an administrative user inter- tailored indexes that can speed up common accesses to digital face and a data import utility for archival repositories. Finally, library sets. we discuss and evaluate a prototype repository we have built, the Stanford Archival Vault, SAV. A second area where