Analysis of Identify TDR Requirements

Update 2008-08-28

All references to 'RLG/NARA audit checklist' in this analysis document is to the 'old' version: rlgnara-repositorieschecklist-V2007.pdf

I have saved a copy of this version in http://merkur/svn/doms/trunk/docs/TDR/rlgnara-repositorieschecklist-V2007.pdf.

The new (2008) version is the TRAC: Criteria and Checklist. We should take time to go through this revised version. We could also take a look at the Platter checklist.

Introduction

In this task we identified three sources of Trusted Digital Repository checklists, referred to as RLG/NARA, Drambora and Nestor. The documents are linked to in the Resources section. We selected two of these documents as the most important. RLG/NARA is a well known list and widely circulated, and DRAMBORA is a certification framework in development by DPE, where The State and University Library is a partner.

From these two documents, we evaluated the checklist of requirements. In RLG/NARA there is an actual checklist, where DRAMBORA contains a list of risks that should be addressed. Note that DRAMBORA calls this list only an example, and a good repository of risks should be developed.

We evaluated each of these requirements/risks with regard to their relevance in DOMS. We then grouped them under headlines relevant for DOMS, and finally we matched each of these groups with the tasks in which they should be addressed.

The result of this work was a few new tasks: TaskC.2, TaskC.3, and TaskD.4. Apart from that, in each task the analysis document has been updated with the groups of Trusted Digital Repository requirements we should address while working on this task.

The rest of this document lists the groups of requirements we identified. This is a condensed version, in the more detailed list the headlines of each requirement/risk is also included. For full details, it will be necessary to cross reference these lists with the source documents.

As for integration with Planets, we did talk with Lars and Thomas, and the results was that we cannot address the exact promises from Planets nor the integration with Planets at this time. TaskC.2 will have to be resolved at a later time, when Planets has progressed a lot further.

RLG/NARA (A-D) AND DRAMBORA (R)

Extracted requirements from the documents referenced below as Resources. See these documents for details.

See also the more detailed list with headlines of each requirement/risk.

Documentation Requirements

These documentation requirements are new, and we should probably update the WBS with new tasks.

Communicate requirements out of scope for DOMS but relevant for SB: A2.*, A3.3, A4.*, A5.*, B1.2, B3.11, C1.*, C2.1, C3.1, C4.*, D1.6, D1.8, D1.9, D2.3, D3.1, D3.3, D3.4, D3.5, D3.6, R01-10, R13-14, R16-47, R57-58, R72, R74-76, R78
Document and communicate repository principles for TDR fulfilling requirements from A1.1, A1.2, A3.4, D1.8, R12, R73
Need documented testing procedures: D1.9
Need documentation of software dependencies: D2.2

IT Maintenance Department Requirements

These requirements also need to be documented and communicated.

Requirements to IT maintenance department: D1.1, D1.2, D1.5, D1.6, D1.7, D1.10, D2.1, D3.1, R30-47, R60, R74-76
Requirements to IT maintenance department which we need to have feedback on: D1.3, D1.4, D3.4, R53, R61

Deployment and Collection Specifics

We should probably take a quick look at these soon and decide which requirements are relevant now and which are relevant later in the project.

Organizational procedures required around deployment A3.*, D1.8, D.2.2
Collection specific requirements we should consider to some degree in DOMS B1.3, B1.4, B1.6, B2.3, C2.1, R72

Ingest

The following should be considered when designing ingest.

File integrity checking during ingest, fail on errors: B1.6, B1.8, B2.6, R48
Proper reporting required from ingest module, ensure only complete objects are ingested and reported ok, report indicates preservation status B1.7, B1.8, B1.9, R48-49, R63
Ensure received data cannot be modified during ingest: R50
Preserve existing PIDs in metadata: B2.5

Repository Contents

The following items are relevant to metadata and digital object design decisions and require documentation too and they should be considered early in the project.

Must be possible to document properties we will preserve for all digital material B1.1, B3.5
Must have preservation strategy for certain classes of digital material B1.1, B3.1, R65, R66
Sufficient metadata about digital objects required: Technical, administrative and preservational as well as descriptive, generated or analyzed to be correct, fail on errors: B.1.5, B1.8, B.3.4, B3.6, B4.1, R11, R54, R62
Format definitions required, note that the intended usage of the format is relevant as well, use standard format registries: B2.1, B2.2, B3.3, R48, R62, R73
Standard permanent UIDs needed, accessible from the rest of world: B2.4, R64

Rights

Must be possible to document copyrights and restrictions on all digital material in a format that it is possible to act on/Require enforced access rights A5.3, B.5.1, C3.3, C3.4, D3.2, D3.3, R15-17, R18?, R36, R52, R76

Repository/Object Integrity

The following items are mostly requirements to logging and should also be considered early in the project.

It must be possible to document how we ensure integrity A3.6
Must be able to document the data we have in the repository is derived from the data we received, and how: B2.3, R51, R54-56, R59, R68
Integrity audit needed - read the RLG/NARA source: B2.7 (+R55? +R56?)
Referential integrity required between file storage and metadata (creation and validation): B4.2, B4.3, D1.5, D1.6, R64, R69

Preservation Requirements

Need preservation strategy (Planets): B3.1, B3.2, B3.5, B3.9, B3.10, R65-67
Must be able to validate/evaluate results of preservation action (Planets): R66-67
Active bit preservation required: B3.7, D1.5, R61
Audit trail required: B3.8, B3.11, D1.6, R70

Access

Data must be searchable/discoverable: R71
Dissemination must always return correct results, and reject requests otherwise : B5.3-6, R77
Access log required: B5.2, C3.2, R36