US 9,812,134 B2
Using bandwidth allocation data to determine a breakdown of VoLTE calls by AMR coding mode
Rajesh Garre, Redmond, WA (US)
Assigned to T-MOBILE USA, INC., Bellevue, WA (US)
Filed by T-Mobile USA, Inc., Bellevue, WA (US)
Filed on Nov. 3, 2015, as Appl. No. 14/931,713.
Prior Publication US 2017/0125021 A1, May 4, 2017
Int. Cl. G10L 19/002 (2013.01); H04W 72/08 (2009.01); H04W 4/16 (2009.01); G10L 19/00 (2013.01)
CPC G10L 19/002 (2013.01) [G10L 19/0019 (2013.01); H04W 4/16 (2013.01); H04W 72/087 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
accessing a first database of call detail records (CDRs) containing information associated with a plurality of communication sessions established over a mobile telephony network;
selecting, from the CDRs, a first set of CDRs that are associated with communication sessions that used a long term evolution (LTE) radio access technology (RAT);
selecting, from the first set of CDRs, a second set of CDRs that are associated with communication sessions having a Quality of Service Class Indicator (QCI) value that corresponds to voice over long term evolution (VoLTE);
selecting, from the second set of CDRs, a third set of CDRs that are associated with communication sessions having durations greater than a threshold period of time;
selecting, from the third set of CDRs, a fourth set of CDRs associated with communication sessions that were released from a bearer of the communication sessions corresponding to the fourth set of CDRs;
selecting, from the fourth set of CDRs, a fifth set of CDRs associated with communication sessions that were allocated an amount of bandwidth greater than a threshold bandwidth;
referencing a second database containing correlations between adaptive multi-rate (AMR) speech coding modes and bandwidth values;
classifying the CDRs of the fifth set of CDRs into groups of different AMR speech coding modes based at least in part on the referencing; and
recommending a capacity planning metric or a resource allocation metric based at least in part on the classifying.