RSUBMIT Performance Issues for 5B row dataset

[tiramisu]

I have a very large transactions table that contains over 5 billion rows, with about 20 columns in the server. 

Because it is so large, I am unable to perform any analysis on this, as the connection the remote server always times out and get the following error:

Spoiler

ERROR: Conversation termination; status=1.

I've tried to create another table to store just data within one year, or even do an inner join to another table to store a subset of customers' transaction details, but neither have worked - I continue getting the time out error...

Has anyone worked with this kind of dataset? Any tips? 

[SASKiwi]

How long does your SAS program run for before it times out? Is the table a SAS table or a database table? If it is a SAS table then what is the size of the sas7bdat file?

[tiramisu]

Usually about 3+ hours. The table belongs to a database.
Below is the exact message I get:

ERROR: Conversation termination; status=1.
ERROR: A link must be established by executing the SIGNON command before you can communicate with
SERVER
ERROR: Remote submit to SERVER canceled.

[SASKiwi]

Check with the database administrator to see if there are any limits on query run times. What database is it? Do you have access to another database query tool to try a test? With such a large table you may need to switch to SQL PASSTHRU to use database-specific performance improvements.