Information Technology Reference
In-Depth Information
(e.g., periodic fixity checks). Avoiding the transfer of the data to the higher applica-
tion not only saves network bandwidth, but also simplifies the LTDP system, which
in turn results in higher overall system reliability.
Although designed and built as the preservation-aware storage component for the
CASPAR project, PDS's flexible layered architecture enables its use as the storage
subsystem in other preservation settings as well. PDS variants have been built that
integrated with an ECM solution, and over a plain file system. These implementa-
tions demonstrate that PDS can extend a preservation-agnostic archival storage to
provide LTDP functionality. Since data subject to long-term data preservation may
already reside in existing systems and archives, easy integration of PDS with other
(existing) systems is important.
The PDS subsystem may be improved and completed in several aspects. To
enhance and complete the support for the CASPAR authenticity model, PDS
should support authenticity protocols explicitly, e.g., by implementing Authenticity
Protocol as an object and preserving each protocol as an AIP. PDS should support
the execution of such a protocol object whether it is a pre-defined protocol imple-
mented in PDS or one loaded and executed by external users. This enhancement
will provide uniform behaviour to internal (automatic) and external (manual) proto-
col executions. The authenticity protocol history will be documented transparently
for all protocols by preserving them as AIPs in the system.
Another aspect that requires additional research and absorption into the PDS
implementation is a placement mechanism that takes into account the different
parameters that influence the optimized clustering of AIPs to be moved to secondary
storage. These parameters involve understanding the relations between AIPs, pre-
diction of access patterns of AIPs, legal issues and aspects related to the physical
secondary storage (e.g., capacity, reliability etc.). In addition, there is a need for a
standardized format that will describe the content of each cluster in order to make it
self-describing and self-contained and thus interpretable by future systems. Towards
that end we are working on Self-contained Information Retention Format (SIRF)
standard in SNIA Long Term Retention working group [
199
].
17.6.5 PDS Component Details
Component name
CASPAR Preservation datastores
Component acronym
PDS
