The project has set up an infrastructure based on data replication (different machines located in different sites).
The current Digital Stacks prototype is now turning into an operational service based on two main deposit sites (managed by the Biblioteca Nazionale Centrale di Firenze and by the Biblioteca Nazionale Centrale di Roma) and a dark archive (managed by the Biblioteca Nazionale Marciana).
Certification to ISO 27001 international security standard will be the basic prerequisite for the selection of a data center. Each institution (Florence, Rome and Venice) will select three different data centers owned and managed by three different companies (to reduce the risk of “domino” effects). Each main site is composed of a set of autonomous and independent nodes. In turn each node on a given site has a mirror node on the other site: the Digital Stacks service does not rely on “master site / mirror site” architecture and each site will contain, in a symmetrical way, both master nodes and mirror nodes. The autonomy of each node offers several advantages from the point of view of scalability (there are no limits to the number of couples of nodes in each different site) and from the efficiency side (the possibility to accept requests for reading or writing is distributed on the different nodes). The elements that compose the nodes are simple and personal computers which can easily be replaced, using widespread and inexpensive SATA technology. The infrastructure does not rely on custom or proprietary software but is based on an open source operating system and utilities (widespread acceptance means less dependencies). Data replication relies on open source disk synchronization utility. Each physical file is replicated twice on different computers within the same node. The dark archive also contains two copies of this file on two different computers. As a result within Digital Stacks each physical file is replicated six times.
Three fundamental requirements are at the basis of the project:
- scalability: the deposit can be extended indefinitely and it can achieve its objectives independently from the quantity of archived data
- reliability: the deposit uses all the strategies necessary to contrast the risks of loss, corruption of data and of uncontrolled access
- economy: the efficiency of the deposit system (in terms of cost - benefit ratio) must be based on reasonable and sustainable costs.
The system architecture, based on the ISO 27001 certification (international standard for an Information security management systems), has taken into account also the DRAMBORA toolkit and the TRAC checklists for the certification of trusted digital repositories.
Further results of the project:
A comparative study of the main software for the management of current digital repositories has been developed. The study has considered only open-source software and the software in use at the Florence National Library. (Read: Lorenzo Mestriner, Confronto fra software per la gestione di Digital Repository, settembre 2006).
Moreover the evaluation of the procedure for the Digital Stacks certification as trusted digital repositories has been analysed. (Read: Ivano Greco, Giovanni Bergamin, DRAMBORA (Digital Repository Audit Method Based on Risk Assessment), giugno 2008).
