You could try using connection pooling from web server side to manage these many users.

You got it. Hard coding provides optimizer a specific value to work with and can look at index value spread and determine a much more accurate estimate of returned row...

PK is a logical design construct used to uniqly identify a tuple of an entity. Physical implementation requires that columns chosen to be PK, should be not null, and have...

You could try "dirty reads" on Subscriber side - it would only work if such arrangement is acceptable to users. Or you could try to do bulk updates only during...

As far as I know there is no set limit on number of transactions a SQL Server allows. It depends on how powerful SQL Server setup is, which depends on...

I'm glad it worked out fine. That 'in process' option is key in making Sybase server working properly as linked server.

First Sybase 10 is way too old - ask your Administrators to upgrade.

I would suggest defining a linked server in SQL Server 2008 pointing to Sybase 10 server, using ODBC...

If you are seeing self blocking for extended period of times - consider setting 'max degree of parallelism' to 1. But to take its affect on processes already running you...

Now a days disk space being plenty and cheap (and faster than tapes), people often use file system for initial backup device, later to be copied to tape. It is...

Unless reporting users are looking for up to the minute or second data, log shipping should suffice. You could also look into mirroring.

That is definitely wild 🙂 A page, either in memory or on disk, can not be shared between different objects.

As to original post - there must be something missing in...