Hi Thomas,
Since you are talking about newspapers, do you pages in terms of millions or
number of newspapers, Also what kind of concurrent user load is expected?
We did couple of implementations and DSpace optimization In past where we need
to manage similar kind of data load I.e. 10 - 25
"Normal" infrastructure issues aside—disk, RAM, CPU, etc—I'd say that
the handle server would be the "Y2K" type issue to be concerned about.
I don't think anyone definitively knows the limit of the handle
server. In May last year I asked about this:
That is a LOT of items! How many requests per day/hour/... do you expect to
have? Is it in the same proportion of items?
I believe that the real problem is the quantity of requests, once the items
will only be another row into the database. But to query over them will be
a real challenge for
I believe that 20 million items would be a record for a DSpace repository.
At Georgetown, we have more than 500,000 items. A majority of these items
contain only metadata which may have simplified the issue for us. We have
found DSpace to scale very nicely so far.
I believe that the University
