on Serials in the Digital Environment
THE LOCKSS PROGRAM
Victoria A. Reich
Libraries build collections to serve communities. They are an
important memory organization. The web, while bringing many benefits
has had a powerful negative impact on libraries: they are no longer
able to easily and affordably build local collections. They lease
feematerials and access free materials. This failure to collect
digital artifacts will create a growing dark ageof our times.
"What is a library anymore anyway?" by Michael
A. Keller, Victoria A. Reich, and Andrew C. Herkovic, First
Monday, volume 8, number 5 (May 2003) at: http://www.firstmonday.org/issues/issue8_5/keller/index.html
LOCKSS (Lots of Copies Keep Stuff Save) was developed to address
this problem http://lockss.stanford.edu. LOCKSS open source software
provides a peer to peer, distributed, persistent access preservation
system for web delivered information. This easy to use, inexpensive
to maintain system will be released for production in early 2004.
Brief description and goals http://lockss.stanford.edu/projectdescbrief.htm
Participating libraries and publishers http://lockss.stanford.edu/projectstatus.htm
Technology: The LOCKSS software turns a personal computer
into a preservation tool by transforming it into persistent web
cache. In 2004 one LOCKSS computer can hold approximately 2,800
years of an electronic journal at an average cost of 35 cents/journal/year.
Given the historic growth of computer storage capacity, we predict
that by 2007 one LOCKSS computer web cache will hold approximately
23,000 years of a journal at an average cost of 7 cents/journal/year.
Libraries provide users access to LOCKSS preserved content by configuring
the LOCKSS cache computer as a transparent proxy on their institutional
Brief technology overview http://lockss.stanford.edu/techoverview.html
Collections: What might librarians collect through the
LOCKSS system? We suggest librarians consider owning rather than
leasing electronic journals accessed through individual institution
subscriptions or consortia licenses. We also suggest librarians
consider coordinating collection efforts to collect and preserve
important freely available electronic titles. Coordinated collected
development efforts are underway for humanities literature http://www.indiana.edu/%7Eletrs/lockss,
government documents http://lockss-docs.stanford.edu/ and other
genres. The LOCKSS technology will provide persistent access to
materials that are delivered through HTTP; published in serial;
have a reasonably stable URL structure; have or can be made to
have an authoritative version.
Collection development http://lockss.stanford.edu/collectionswork.html
Publisher must grant permission to the LOCKSS crawler before
the system will collect copyrighted content. Permission must be
granted for every individual titles archival unit(usually a volume)
via a LOCKSS Publisher manifest web page. This manifest page can
link to and bundle together content not routinely associated with
the electronic journal (front matter, editorial instructions, and
for example, the titles CONSER bibliographic record) for long term
preservation in the LOCKSS system.
Publisher actions http://lockss.stanford.edu/publisheractions.html
Summary: LOCKSS is a working example of a distributed repository.
We believe the major threat to successful long term preservation
is budget cuts. A major advantage of a distributed repository model
is that the cost of preserving content is shared among many participants.
There is no central point of failure, neither technical nor budgetary.