Re: [Serusers] Creating CDRs from SERs accouning records

If I am correct I believe CDRTool uses mediaproxy to prevent the phantom BYE record. Adrian can correct me if I read wrong. Invites without BYE's are common No one answered the phone. We just Log and drop records that dont have matching INVITE,ACK, and BYE. The reason as I have seen for the missing BYE (discussed alot on here) is the UA or the Gateway dont send a BYE for whatever reason. We havent seen that many , infact not one yet knock on wood. We thought about just estimating and calling them one minute calls. The problem we are having is we route some wholesale traffic thru us via trusted IP and account for it (Easy Part) but we also get DID's from that Company so all traffic comes from same IP. DID's work fine but it broke our parser. LOL was charging them for our incoming calls OOPS!!! I guess for now I am going to move the wholesale Trusted IP stuff to its own box. Eric eric(a)rackspeed.net ----- Original Message ----- From: "Frank Fischer" <frank.fischer(a)digitalnomads.ch> To: "'Dave'" <ddx66(a)yahoo.com> Cc: <serusers(a)lists.iptel.org> Sent: Wednesday, November 30, 2005 5:14 PM Subject: RE: [Serusers] Creating CDRs from SERs accouning records > > Do i get this right - the matching is just done by searching a matching > INVITE for each BYE where matching means that the callid must be the > same? > Anf of course by calculating the callduration, normalize certain > fields and > "grap" phonenumbers out of the sip uri? > Is it really that "simple"? Since, on one hand side, i read about cases > where there is no BYE at the end of a call (i think it had to do with > mediaproxy), on the other hand side, when i look at "my" acc table, i > can > find some records, where there are multiple BYEs with the same callid or > INVITEs without a BYEs. > > I understand, that missing BYEs primarly are a problem of the > accounting, > not of the normalization, but, how ever, i'm quite sure that i won't be > possible to have a 100% accurate accounting so i thought that there > might be > the need/possibility to do some "fixing" during normalization. I.e. i > think > i understood that the commercial version of CDRTool does so for missing > BYEs? > > Am i completely wrong? >