Experimental Physics and
Industrial Control System

Timo Korhonen via Core-talk <core-talk at aps.anl.gov> · Mon, 25 Oct 2021 08:59:17 +0000

Thanks Dirk,

These are good hints. I will get back when and if we can understand the issue better.

Timo

On 2021-10-25, 10:54, "Core-talk on behalf of Zimoch Dirk (PSI) via Core-talk" <core-talk-bounces at aps.anl.gov on behalf of core-talk at aps.anl.gov> wrote:

    Yes, we do something similar at PSI:

    Beamline Network -> gateway A -> Machine network -> gateway B -> office network.

    I do not see such a delay.

    We have seen load related effects. In particular loading the gateway with arrays "freezes" the monitors on scalar
    values. Recent observations with camera IOCs suggest that this is not gateway specific. Somewhere in CA monitors on
    arrays seem to get higher priority than scalars. No idea where.

    Try to sniff network traffic in the middle. Do you see the search requests coming through? Do you see the replies?
    Maybe your routers/firewalls are dropping UPD packages? (Which is allowed. UDP does not guarantee delivery.) We had some
    problems when updating our Cisco routers and it took about a year to get it fixed by Cisco. But I do not know details.

    Dirk

    On Mon, 2021-10-25 at 08:35 +0000, Timo Korhonen via Core-talk wrote:
    > Hi,
    >  
    > We at ESS are using CA gateways in a “chain” configuration, i.e., roughly in the following way:
    > Machine (aka Technical) network -> gateway A-> internal DMZ -> gateway B-> office network.
    >  
    > Does any of you have any experience with such a configuration?
    >  
    > What we have been seeing are somewhat oddly constant access delays, for instance caget taking more or less exactly one
    > second to read an otherwise idle PV via gateway B.
    >  
    > Not surprisingly, this appears to be related to the traffic on the networks (i.e., machine operation.)
    >  
    > We have not yet done any systematic studies on this, I just thought to ask if anybody has experience or thoughts about
    > this kind of issue.
    > Is this usage pattern something that should rather be avoided?
    >  
    > Cheers,
    >  
    > Timo
    >  

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System