Dear Steve,

>> At the moment we are ingesting a large collection of digitized  
>> documents (21.000 representations with an average of 80 pages  
>> each). More than 14.000 representations have already been ingested  
>> and the performance of the repository has not been affected.
> What is the cost of ingesting this content into RODA, in terms of  
> time or money? Can you give a broad lifecycle estimate of costs as  
> well?

That is a really difficult question. To be honest I have no idea what  
the associated costs of ingesting large quantities of information will  
be. Several costs are involved:
- Manpower at the producer to prepare and send SIPs to RODA (assuming  
that the producer doesn't have a document management system that  
integrates nicely with RODA)
- Computing power during automatic validation stages, migration  
procedures, metadata extraction and PREMIS generation, etc.
- Manpower at the National Archives to semantically validate each  
submitted SIP
- Storage costs
- Electricity, cooling, etc...
- etc...

The list goes on and on... so at this time is really hard to make  
predictions. I'm sure the people at the National Archives will have a  
clearer idea on that as they will support the repository in the long  
run.

>>
>> RODA provides two distinct interfaces, a Web based user interface  
>> and a Web services API. Integrating RODA with other systems is  
>> straightforward as long as they support Web services. In what  
>> respects OAI-PMH, a little bit of additional work is necessary to  
>> make it compatible with the protocol.
> I'm interested in the architecture of repository preservation  
> systems. In this scenario you suggest the possibility of harvesting  
> from a repository into RODA? Rather than adding RODA-like  
> functionality to the repository? I realise this part is beyond your  
> actual application, but I'm interested in your thoughts.

RODA was not designed to harvest information from other systems. But  
it was designed to implement a set Web services that allow any other  
system to submit new information into RODA. Additionally, these  
systems may query the repository, download information (i.e., DIPs),  
etc... Every possible action in the repository is open for use for any  
system as long as they use the adequate access credentials.

Please take a look at all the possible actions here http://roda.di.uminho.pt/roda-core/services