Re: SQL*Net V2 and performance?

From: Roderick Manalac <rmanalac_at_oracle.com>
Date: 2 Oct 1994 07:47:19 GMT
Message-ID: <36loi7$4jf_at_dcsun4.us.oracle.com>


davidsonj_at_gtewd.mtv.gtegsc.com writes:
|> Does using V2 of SQL*NET and the MTS capability buy any performance
|> improvements?
Difficult to say. It depends on the environment. It mainly is designed to allow a larger number of concurrent connections. Efficiency of server use may increase if have applications running involve alot of think time on the user end. However, individual response times would be slower since messages have to pass from the client through the dispatcher, through the SGA, and finally to an available shared server.

|> I heard about something called UPIALL (sp?) that reduces the number of
|> packets sent across the net, but if I'm using stored procedures already
|> and am using the delayed binding option, will this be of any benefit?
By default most recent versions of Oracle products already use the new two-task protocol to talk to the server. You only have to worry about how to use it if you are writing OCI programs. Using stored procedures is a good thing nonetheless.

|> We have around 30 sessions (maybe 40 with SQL*plus sessions running).
|> Is that enough to benefit from the MTS capabilty gained with v2 of
|> SQL*Net?
If your machine would have otherwise been lacking for resources, then yes. While it may not be necessary to use MTS, it is still a good idea in the long run to use SQL*Net V2 instead of V1 if possible.

Roderick Manalac
Oracle Corporation Received on Sun Oct 02 1994 - 08:47:19 CET

Original text of this message