Re: [Serusers] Creating CDRs from SERs accouning records

It is usually enough to match INVITE with --first-- of recorded BYEs with the same CallID. It is fairly accurate. From my experience I would say every month 1 or 2 calls have missing BYEs. You give them "1 minute" duration or just forget ;-) BTW this is just a top of an iceberg somewhere in front of us. And it does not only concern billing. It concerns management, provisioning, terminal capabilities, charging, multimedia and so on. We surf towards IMS-like architectures, wherever we want or not. Look at today - we create our own billing systems where every one is different, proprietary and acting on different requirements. IMHO we should rather work together, wherever you call it OpenSource or not. Let's start with billing. Let's define common requirements, like: 1) tarrification policy (peak/offpeak,initial,post-call charges and so on) 2) way to manage endprices,margins,currencies 3) different database versions to let everybody choose his backend: sql or directory database 4) common data interchange format (I propose IPDR, VoIP profile) 5) common communication protocol (I propose SOAP) We should not work on payment management, as BSS systems should do it for us, provided that nice data is exported from SER. But if someone wants - OK, payment management may be done as well. Some may say that SER is not a place for SOAP, IPDR and related stuff. But let's be honest - world is moving into similar direction and technologies such as Parlay X Web Services, JSLEE or IMS are just few examples. There is also ongoing effort to integrate network services around directory databases, wherever it is JNDI (in JSLEE) or LDAP (in OpenSource projects). I was working for most of this stuff, with success, but a little outside of SER. SER was always somewhere there, but with MySQL stripped almost completely (in fact only 'location' table is used as a kind of HLR). And I see from private emails or contributions that many (all?) of us have done the same. People are using LDAP backend with RADIUS (worked great for us) or without RADIUS (directly to LDAP, worked great too). I have got reports of people having customer base of 5 milion and looking for 10k transactions/second with LDAP backend. In fact this is OK - you just choose those parts of OpenSource project that fit your needs and integrate. Thats what OpenSource is for. Tens of great people, like SER core developers are working on great software that SER is now, they work for free and make this software better. But so many users are going towards the 'next big thing'. At the same time tens of other people are working on proprietary platforms for billing, LCR, Web Services or even IMS in their companies. It is OK, but why not provide them a good '--next killer platform--' and let them concentrate on creating better --services --? Consider this quotation from SLEE primer available at opencloud.com: "We must, as an industry, make it abundantly clear to all commentators that we are in business to create the most potent and fundamental part of the social fabric of the 21st century. (...) Operators must recognise (...) not just a single service, like video or gaming. By implication, therefore, (...) operators should be creating a killer environment." Copyright Nicolas Foggin, Director of Strategy & Futurology, Orange SA Copyright Alcatel Telecommunications Review, 2002 You see? Provide killer environment. Let's start with billing. -- Arek Bekiersz Eric Haskins wrote: