Vlad,

Thanks for your suggestions. I will look into them. I am also going to try keeping the DataTables that I am using to store the report on the client smaller, saving more often and then creating new tables. Maintaining the tables of calculated data has also got to drag the client down.

We don't really want to run a second app on the database because it is a pretty busy machine with a lot of people querying it and we need to keep it focused on serving up queries. The only way that I can imagine serializing on the client would help would be if Windows and .NET together do a really bad job of maintaining large DataTables, doing a lot of swapping from RAM to harddisk.

Thanks again.

>Hi David,
>
>I got a few questions and few suggestions:
>
>Do you need to run the report on the client machine? Thats a huge report...300000 records.
>I was thinking that you may be able to run the process on the server, dump the data into a report table and , when its done, send the formatted data to the client, ready for the report. It still wont be very fast but definately faster then running the whole show on the client.
>
>As far as serializing a dataset as big as the one you are sending to client, .net serialization performance will degrade big time. I suggest you take a look at a Microsoft freebie called DataSetSurrogate (search msdn for) that does a better serialization of large datasets. This, along with compressing and streaming the dataset would work perfectly. You can use Xceed for compressing and streaming ( Thanks Dave Foderick for the pointer:)
>
>Im currently working on a smart client app that downloads data(datasets) and I can load 600000 records very fast( via remoting). I m using DataSetSurrogate, Xceed for compression and streaming and remoting.
>
>As Thomas Ganss suggested, msde might be another option here for dumping those 300000 records and processing the report on the client.
>
>Hope it helps,
>
>Vlad