Hello, <BR><BR>I&#039;m am building an application in .NET on SQL2k that could concievably have 10-20k simultaneous users, meaning at the exact same moment in time 20k people could be executing the same stored procedure. Now, this sproc will be quering a table that could be ~ 100k records, but immediately closes the connection to the db after execution. On to my question. <BR><BR>Would it make sense to have the core of my application in a single database and then break out the one 100k record table into several different datases? That way my 20k simultaneous users would be hitting say, 5 or 10 different databases and each time the individual sproc ran it would only have to query a table with 10-20k records? I would then use distributed queries to grab any other data I needed from the "master" database. My biggest worry or need is to know how many simultaneous open db connections is acceptable on a single database (if only for a millisecond) as well as any performance issues with distributed queries. <BR><BR>I&#039;ve seen posts of people querying tables with 18 million records, so I&#039;m not sure if my setup is something to be worried about, but before fully developing this application I have to be sure my design is scalable. <BR><BR>Any comments on this would be greatly appreciated,<BR><BR>])ry