|Date: ||Mon, 2 Apr 2007 06:38:31 -0400|
|Reply-To: ||"Richard A. DeVenezia" <rdevenezia@WILDBLUE.NET>|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|From: ||"Richard A. DeVenezia" <rdevenezia@WILDBLUE.NET>|
|Subject: ||Re: Can the same box be both Client and Server?|
> I am investigating the conversion / migration of our SAS environment.
> We intend to move from a Client Server Model (PC to Unix) to a MS
> Windows Server Based environment.
Never worked with a clientless server :)
> We have quite a lot of client/server code using remote submits, proc
> download, sysrput and syslput, remote library sevices etc. Obviously
> we need to modify this code to run in the new environment.
Do you plan on _not_ using RLS with the new box ? Probably not wise and
maybe infeasible if you have a concurrent multi-user situation.
> My question is:
> Can I define our new Windows based Server as both the client and
> server in SAS?
> This would allow the code to run with minimal changes.
> We could then convert to the Server-based sas code at a more
> leisurely pace.
> Any other ideas? Comments?
I don't see an obvious need to modify 'this code', unless the code relies
heavily on shelling out to the Unix environment.
The SAS multi-vendor architecture (MVA) normalizes client/server
implementations to the point of astonishment. If you name and address your
new MS box the same as the old Unix box, you might be surprised at how
little has to be changed. If you can't do that, and you have reams of
legacy ad-hoc code with hardcoded connect values, you will want to find a
good search and replace tool.
- How are your information consuming clients going to connect to the SAS
services running on the MS box ?
- Have you performed benchmarks that suggest your C/S implementation under
MSWS will not be responsive enough?
Richard A. DeVenezia