Hi folks; I posted this to a different group and recieved no repsonse,
so thouth I would try here.

I'm looking for some advise about handling large datasets
generated for my Windows forms app.
The datasets returned have grown very large as new items are
continuosly inserted, and the time it takes to load those datasets in
turn, is taking several minutes.

The question I have: is there a better approach to loading datasets
from SQL Server.

The design uses a one to many relationship on two tables, with the
child table holding many rows of data and continously being updated.
However, the original "load dataset" command ultimately loads all
parent records and child records into memory, then the UI allows the
end user to select a parent table with its related child table for
editing.

This has worked OK for a while, but because these tables have grown
large, I'm curious if there is a way to do this with out A: making a
lot of round trips to the server over the network, as may be required
using SPROCS; or B: loading the entire dataset into the client machine
memory.

Currently I'm developing a framework to use with an app server, and
intend to extend the application to several desktops, and also plan on
adding an additional physical layer for such. My thinking is that
moving the business logic and rules to the app server could help reduce
the latency experienced by the end user.

Any ideas, links or suggestions would be most appreciated!

TIA!