map/reduce is used in nutch to do the actual processing of the index. the
searching (was when i last looked) a daemon which could run on mutliple
servers. looking at their solution for this is useful. I have done such a
thing myself before a while back.
a branch utilising boost-mpi exists for cluc
Thanks Ben for quick reply.
I read somewhere , Nutch uses mapper and reducer to do so...but really donno
how to go abt it.
On net I am not able to find gud papers related to it.
If you could provide me the links where I can find more detail abt it.That
would be helpful
has parallel multi sear
First of all have a look at the mpi branch ofclucene. It achieves the
distributed searches.
Nutch is a project which I know does this sort of thing u r looking
at, so I look at the theories of that project may give u some good
ideas.
Good luck and please let us knowhow you go
ben
On Thursday, S
