Experimental Physics and
Industrial Control System

Benjamin Franksen <[email protected]> · Wed, 13 Dec 2000 22:32:56 +0100

Ralph Lange wrote:
> 
>   > The *client* has the final decision, which pvs to group into a request,
>   > subject to the restriction that these pvs must be provided by the same
>   > server instance.
> 
> Forbidden restriction, as it breaks the CA axiom that connections are
> done by channel name only, indepent from servers.
> 
>   > New record types will therefore *not* break existing clients.
> 
> But splitting a database on two servers will. Not a good idea.

I see. Scrap the idea to combine pvs from more than one record. It was
something I thought would be nice to gain, while we are at it. What
about the rest?

As I think about it: The restriction is not necessary. The client
combines only those pvs into an actual CA request that (momentarily)
come from the same server or alternatively those which belong to the
same record. It will create a subgroup for each server resp. record and
this subgroup will be packed into a combined request.

Re-grouping can be done automatically as soon as one of the connections
break or come up again. This would be completely transparent for the
client tool.

Ben

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System