" /> Ping request time out setting in SIP - Genesys CTI User Forum

Author Topic: Ping request time out setting in SIP  (Read 2835 times)

Offline aditya.bhattacharya

  • Newbie
  • *
  • Posts: 2
  • Karma: 0
Ping request time out setting in SIP
« on: December 23, 2015, 11:51:18 AM »
Advertisement
We have an inbound SIP solution with geographically located Contact Centers connected by MPLS with enough bandwidth. When there are packet drops in MPLS network between DC and Call Center, voice issue occurs. With 2-3 consecutive "Request Timed Out" voice starts breaking/cranking, but with more than 3-4 consecutive "Request Timed Out" the call disconnects. I can understand that RTP flow breaks with multiple RTOs and calls will disconnect. But I want to know if there is any kvp/setting in SIP/SipEndPoint (we use custom SipEndPoint with GAD) where we can control this network packet drop parameter and keep the RTP session alive for few more seconds until far end starts responding.

More information can be furnished upon request.

Offline Kubig

  • Hero Member
  • *****
  • Posts: 2755
  • Karma: 44
Re: Ping request time out setting in SIP
« Reply #1 on: December 28, 2015, 07:33:58 AM »
I think, there are no option how to achieve that on genesys level, but you can try contact Genesys directly to ask them.

Offline genesysguru

  • Sr. Member
  • ****
  • Posts: 293
  • Karma: 12
    • Genesys Guru Blog
Re: Ping request time out setting in SIP
« Reply #2 on: December 29, 2015, 11:49:26 AM »
Hi,

As a starting point I would ensure that RTP silence supression is not configured to ensure that the RTP stream carries on in periods of audio silence and hence keeps the network path open.

Regards
Craig

Offline aditya.bhattacharya

  • Newbie
  • *
  • Posts: 2
  • Karma: 0
Re: Ping request time out setting in SIP
« Reply #3 on: January 04, 2016, 11:12:45 AM »
We conducted a series of tests both in Test Env and Production. With no changes (like Silence Suppression) in environment we made a test call to an agent and disconnected the LAN interface and waited for significant number of RTOs. All this while we kept the customer call alive but no voice interaction with agent. Soon GAD displayed "Connection to server is down". But as soon as we reconnected the LAN interface on agent system and IP was reassigned, the customer call was re-established successfully with voice.

This further perplexed us with an indication that there is perhaps no problem within Genesys, but there is something within the MPLS/Router/Switch(L3/L2) which may be releasing the RTP session during packet drops/RTOs or it may also be possible that customers are genuinely disconnecting upon getting silence on call.

Is there is any config on Router/Switch which handles RTP session/port and is responsible for call release in this scenario?