Preservation DataStores: New storage paradigm for preservation environments

Abstract

As the world becomes digital, we are in ever greater danger of losing business, scientific, artistic, cultural, and personal assets. The threat of such a digital dark age stems from the fact thatunlike physical records that may survive decades, centuries, or even longer without advanced planning-digital records will not survive without planning and diligence. Everything needed to keep digital records viable will become obsolete, including hardware, software, processes, and formats. Consequently, digital preservation environments are needed to ensure the ability to access valuable digital records decades from now and, more significantly, to ensure the interpretability of the records once accessed. We describe Preservation DataStores, an innovative storage architecture that facilitates robust and optimized preservation environments. It is a layered architecture that builds upon open standards, including Open Archival Information System, XAM (Extensible Access Method), and Object-based Storage Device. We also describe the integration of Preservation DataStores with existing file systems and archives and discuss some design and implementation issues. We are developing Preservation DataStores as an infrastructure component of the European Union CASPAR (Cultural, Artistic and Scientific knowledge for Preservation, Access and Retrieval) project, where it will be used to preserve scientific, cultural, and artistic data.

PDS Cloud: Long Term Digital Preservation in the Cloud

“…The foundations of PDS Cloud were established in PDS [4], a preservation storage architecture using Object Storage Devices (OSD). Here the scope is expanded and adapted for the cloud environment.…”

“…It is much more efficient, cheaper and more reliable to do these tasks near the data, namely instead of moving the data from storage to a processing machine, move the computation module to the storage server and run it there. Computational modules deployable in storage are called "storlets" [4]. We plan to develop extensions for open source storage clouds, such as Openstack Swift, to support storlets.…”

A Cloud Environment for Data-intensive Storage Services

“…These principles are reflected in our architecture through a programming model for these computational agents, called storlets, which were originally introduced in the context of digital preservation repositories [24]. Storlets are released into the cloud and activated by events on data; they define not only the computation, but also triggering conditions whereby storlets are activated (e.g.…”

Towards SIRF

“…The papers examine how existing systems address these concerns and suggest methods to ensure secured long term survivability. Preservation DataStores (PDS) [10,11] is storage architecture for OAIS-based preservation-aware storage that focuses on supporting logical preservation. It is motivated by the notion that digital preservation systems will be more robust and have lower probability for data corruption or loss if they offload preservation-related functionality to the storage layer.…”