[SR-Users] NatHelper - Received adding on Contact header

Daniel Tryba d.tryba at pocos.nl
Mon Oct 3 15:19:01 CEST 2016


On Mon, Oct 03, 2016 at 03:35:59AM -0700, ycaner wrote:
> it is simple way to solve this and it is softphone product problem. If i did
> configuration as you said , it consumes so much socket , CPU and etc. in
> future.

I can't imagine this will be a significant amount of cpu/traffic. I can
only speak for my users, but near 100% is behind NAT or a statefull
firewall. Treating the users as always NATed avoids a lot of headaches
so far.
 
> I wonder that is there a something about it  on ietf or protocol flows. I
> think it is wrong. 
> 
> What do you think about it.

I can't find any info about this learning process via the Contact
header. But I have seen this kind of behavior with Patton SmartNode
equipment, in these appliances this learning is related to the UACs
keep-alive mechanism. This further supported my idea to never trust any
endpoint device :)




More information about the sr-users mailing list