Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Fri, 12 Aug 2016 15:10:18 -0500

Hi Dehong,

On 08/12/2016 02:48 PM, Zhang, Dehong wrote:
> Is it possible that the channel access server may send duplicates,
> or re-transmit the same messages?  Specially when congested?

The CA server gets informed of value updates by the data source (a
database record for an IOC server, some other software for a CAS server)
and queues each update to be forwarded over TCP once to each subscribed
client. If there is congestion some individual value updates may be
dropped from the queue (and since different clients will accept data at
slightly different rates different updates may be dropped), but the
client should never see duplicates.

Note that if a channel sends a series of updates with values A, B, A, B
and one of the B values gets dropped it may appear that an A was
duplicated, but that wasn't what actually happened. You may be able to
use time-stamps to distinguish that case, providing the source of time
increments fast enough that each value update gets a different
time-stamp; a VxWorks IOC that just relies on the 60 Hz tick interrupt
for its clock probably won't be fast enough for that.

The CA protocol never sends value updates over UDP, so Michael's point
that UDP packets can be duplicated by the network isn't really relevant
here (it does apply to CA name searches and beacons, but those are
hidden from the client application by the CA library anyway).

HTH,

- Andrew

-- 
Arguing for surveillance because you have nothing to hide is no
different than making the claim, "I don't care about freedom of
speech because I have nothing to say." -- Edward Snowdon

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System