Re:2014: The year the Americans broke the internet
by aidenn0 on Tuesday February 25, 2014 @06:55PM
(#7301663)
Note that I've seen VOIP latency increase in the wake of a cable-company deploying a phone option. Disguising the nature of the traffic caused the latency to drop again.
[ Parent
| on HN ]
Re:2014: The year the Americans broke the internet
by rahimnathwani on Tuesday February 25, 2014 @07:15PM
(#7301756)
How do you disguise the nature of the traffic without also changing the route taken by the packets?
I would be inclined to test this with/without a VPN connection, but then I wouldn't know whether the improvement was due to avoiding throttling/whatever, or just that my VPN server has better quality routes to the VoIP endpoint.
I guess if you VPN to the VoIP server itself, then it would be a fair test.
Re:2014: The year the Americans broke the internet
by aidenn0 on Tuesday February 25, 2014 @07:22PM
(#7301805)
All RTP seemed to exhibit the issue, so we just setup a test with two RTP endpoints. We wrote a very simple UDP mangler/demangler (XOR with a fixed key) and ran it on both ends of the RTP connection.