This study presents and evaluates new caching algorithms
for sequentially accessed data that is being concurrently
used by several processes or jobs. This type
of access is common in many data processing environments.
In this paper, we focus on batch processing,
which is often dominated by concurrent sequential access
to data. Several of our algorithms are applicable
to other domains in which sequential scanning of data
is common, such aa query processing or long running
transactions in database systems.