For my parallel searching experiment, I created a federating database
that forwards a search on to multiple databases and accumulates the
responses. It works for searchRetrieve and scan. It doesn't support
sort across the databases yet.
It uses a very simple database model, assuming that all the databases
are identical. So, it does no query mapping or record syntax
normalization. It uses the explain record from the first database on
its list as its own explain record. All of those assumptions were valid
in my experiment where I took a large database and broke it into smaller
pieces.
Ralph
> -----Original Message-----
> From: SRU (Search and Retrieve Via URL) Implementors
[mailto:[log in to unmask]]
> On Behalf Of Eric Lease Morgan
> Sent: Wednesday, October 26, 2005 9:37 AM
> To: [log in to unmask]
> Subject: federated search
>
> What are some techniques y'all would suggest I employ if I wanted to
> implement federated search against my growing collection of SRU-
> accessible indexes?
>
> The primary purpose of SRU seems to be to provide a standard, XML-
> based interface to Internet-accessible indexes. Yet, with my growing
> collection of SRU-accessible indexes, and considering people's
> expectations for searching multiple indexes simultaneously, it might
> behoove me to implement a federated interface to my indexes. How
> might I go about implementing such a thing?
>
> --
> Eric Lease Morgan
> University Libraries of Notre Dame
|