Example video codec of interest/observation - Cisco Spark Room Kit
Software version of sample codec - ce-9.5.0-fb56c25334b-2018-09-13
API request of interest - xStatus MediaChannels Call
Clarification needed -
-In the xstatus mediachannels API request we retrieve jitter, max jitter, loss, packets
-Capturing/storing these datapoints historically for inbound/outbound video and audio channels
-Assumed that packets (xStatus MediaChannels Call [n] Channel [n] NetStat Packets) represents total number of packets sent by the codec (egress) or total number of packets intended for the codec (ingress)
-In this assumption - if there were loss packets we could derive the loss percentage simply as loss divided by packets
-Recently on a highly degraded call we observed that loss percentage - using above formula - was well above 100%. Bringing into question the formula used.
-Based on this recent experience the question becomes - is packet count as revealed in the API request truly total number of packets (successfully sent/received packets and loss totaled) or instead does packet count only represent successfully sent/received packets. If the latter - it would appear that our formula for loss percentage calculation should become (loss) divided by (loss plus packets)?
-Additionally raises question - does this calculation need to be different for inbound and outbound channels? Such as - packet count for outbound channels would be total packets sent by the codec (which would include packets that were lost) meaning formula for loss percentage is simply loss/packets but for inbound channels the formula should be (loss/packets) and for outbound channels the formula should be (loss/(packets + loss))?