- FreeSWITCH-dev - FreeSWITCH Mailing Lists

Implementing Forward-{All, Busy, NoAnswer}, ring policies and DND in FS
by Mathieu Parent 14 May '11

14 May '11

Hello, *Context*: --------- I want to implement the following in mod_skinny (FS-3048 and FS-3047) in a way shared with other endpoints modules: - call forwarding[cfw], - DND and ring policies *Call forwarding*: --------- There are three main kinds of forwarding: - forward all (immediately) - forward when line is busy - forward when no answer We should also store for each the state of the forwarding (enabled or not). I think a reasonable default here would be: - to forward "no answer" to voicemail (and enabled) - to forward busy to voicemail (but disabled) - to forward all to "" (and disabled) We also should define what a busy line is, especially on shared lines (perhaps define a busy threshold, defaulting to 1). Call forwarding should be implemented in dialplan, but endpoint modules should know when a call is forwarded (the bridge event is probably sufficient). *Ring policies and DND*: --------- A do-not-disturb flag can be set on an extension so that the phone doesn't ring (but the user still knows he/she is called). DND is a specific case of ring policies. I propose the following default ring policies: - on idle: ring, blink (toggled to blink with DND button) - on busy: blink only (also available: ring once) This should be implemented per endpoint module. *Proposal*: --------- I propose to use hashes to store the various data: <action application="hash" data="insert/${realm}/${extension}/${data}"/> Where ${realm} is one of: - ${domain_name}-forward-{all,busy,noanswer}-destination: a destination number - ${domain_name}-forward-{all,busy,noanswer}-status: true/false - ${domain_name}-busy-threshold: [0-9]+ - ${domain_name}-ring-on-idle: true/false/once - ${domain_name}-blink-on-idle: true/false - ${domain_name}-ring-on-busy: true/false/once - ${domain_name}-blink-on-busy: true/false The dialplan should be enhanced to manage forwarding. The endpoint modules should be enhanced to manage forwarding (notify that the call has been forwarded) and ring policies. Does this makes sense? How can we improve it? *Next steps*: --------- - proposal discussion - forward-* dialplan implementation (I need help here, I am not an dialplan expert) - mod_skinny forward-* implementation FS-3048 - mod_skinny ring policies implementation FS-3047 - other endpoints implementation Regards Mathieu [cfw]: http://en.wikipedia.org/wiki/Call_forwarding [dnd]: http://en.wikipedia.org/wiki/Do_Not_Disturb_%28telecommunications%29

3 4

Can UUID's change?
by Mark Sobkow 14 May '11

14 May '11

Here's our scenario: The operator logs in to our system, and has their UUID parked. The system calls the customer when requested, which creates a UUID. The operator and customer UUIDs are bridged. We watch for events from the customer UUID to determine when the customer has hung up their end of the call, which causes the operator to be parked again. However, I'm seeing some unrecognized UUIDs after bridging the call (i.e. that weren't returned by the place call code.) Are there any events which might indicate that a UUID is being changed by the system, and that I should be watching for a new UUID instead? -- Mark Sobkow Senior Developer MarkeTel Multi-Line Dialing Systems LTD. 428 Victoria Ave Regina, SK S4N-0P6 Toll-Free: 800-289-8616-X533 Local: 306-359-6893-X533 Fax: 306-359-6879 Email: m.sobkow(a)marketelsystems.com Web: http://www.marketelsystems.com Visit our Blog for industry related information. http://marketel-systems.blogspot.com/

3 2

MEDIA_BUG_START
by Mark Sobkow 13 May '11

13 May '11

I'm getting a MEDIA_BUG_START event. I think there's a correlation between this event and some dropped calls we're experiencing. I haven't found any way to force the event to occur, so I haven't been able to capture a log file of it happening in fs_cli. Our FreeSwitch box is configured to use a SIP Trunk provided by our production Asterisk box, with T1 lines from thereon out. For the most part it works, but one user in particular has a serious problem with dropped calls and I'm trying to figure it out. -- Mark Sobkow Senior Developer MarkeTel Multi-Line Dialing Systems LTD. 428 Victoria Ave Regina, SK S4N-0P6 Toll-Free: 800-289-8616-X533 Local: 306-359-6893-X533 Fax: 306-359-6879 Email: m.sobkow(a)marketelsystems.com Web: http://www.marketelsystems.com Visit our Blog for industry related information. http://marketel-systems.blogspot.com/

2 1

start_dtmf detects dtmf as may times per dtmf as many times start_dtmf issued.
by Anton VG 13 May '11

13 May '11

Just noted that my ESL get as may DTMF events per single keypress as, many times I issued start_dtmf on the uuid. it means if I do: execute('start_dtmf') execute('start_dtmf') I get the same DTMF event twice. Is that expected behavior? Logically, it supposed to to trigger detection on/off

3 5

Re: [Freeswitch-dev] is there any info abount which git tag/ Freeswitch release is stable?
by Tom Parrott 12 May '11

12 May '11

Hi, Does the file at http://latest.freeswitch.org change each day to represent the current state of HEAD? If so I could version that as 1.0.7-yyyymmdd to make it more suitable for RPMs. Thanks Tom >> >> I believe the developers have been discussing making a 1.0 stable branch >> soon with only bugfix and stability improvements and moving new >> development >> onto a new 1.2 branch. If that happens then perhaps you'll be more >> comfortable with that branch. Not sure when that's planned for though. > > > The issue by the way is with developers. While there's a lot of developers > and plenty of people using FS in the wild testing and making bug reports > on > Jira and plenty of people volunteering support on the list and IRC, > something like 90% of the code is still written by half a dozen people. > It's > damn good code though. :) That's too much to make it easy for them to both > maintain a stable branch and work on the newer branch, so up to this point > it has been better to focus on making FS complete, feature rich and stable > all in the same branch. Once the branches diverge, someone is going to > have > to both handle bug reports for 2 different codebases and also look at all > 1.2 changes to decide whether any are work backporting to 1.0 which won't > always be trivial if the code diverges significantly. > > -Steve > > > On 12 May 2011 08:45, Steven Ayre <steveayre(a)gmail.com> wrote: > >> The 1.07 version available at http://latest.freeswitch.org seems to be >> rebuilt >>> each night, does this mean that a 1.07 built yesterday is different >>> from >>> one built today? >> >> >> Yes. Think of it as a release candidate for 1.0.8 >> >> The git head is pretty much the most stable. Development tends to move >> too >> fast to make more regular 'stable' releases. You should of course test >> releases work as expected before putting them into production. >> >> If you find a Git version that you're comfortable with, then you can >> roll >> out that version across your whole network even if there's a newer Git >> head >> by running: >> > git clone git://git.freeswitch.org/freeswitch.git >> > cd freeswitch.git >> > git checkout GIT_HASH >> >> I believe the developers have been discussing making a 1.0 stable branch >> soon with only bugfix and stability improvements and moving new >> development >> onto a new 1.2 branch. If that happens then perhaps you'll be more >> comfortable with that branch. Not sure when that's planned for though. >> >> -Steve >> >> >> On 10 May 2011 19:39, Tom Parrott <tomp(a)tomp.co.uk> wrote: >> >>> Hi, >>> >>> I work for a company selling call tracking services, and we have >>> recently put into production a 3 node Freeswitch cluster to use >>> features >>> such as call recording and call announce. >>> >>> We are using the event socket feature to control freeswitch from our >>> custom written service. >>> >>> I have found Freeswitch a pleasure to use, and much easier to get along >>> with than Asterisk. >>> >>> However one thing that has concerned me is the lack of regular >>> point-in-time releases. >>> >>> When putting services in production, we use RPMs to allow us to easily >>> upgrade (and if need be rollback) software versions. >>> >>> The 1.07 version available at http://latest.freeswitch.org seems to be >>> rebuilt each night, does this mean that a 1.07 built yesterday is >>> different from one built today? >>> >>> If I can offer any assistance with the automation of point-in-time >>> tagged releases I would be happy to help, as I feel this would make >>> Freeswitch more attractive to commercial operations as it would easily >>> fit into well defined release processes. >>> >>> Thanks >>> Tom >>> >>> _______________________________________________ >>> FreeSWITCH-dev mailing list >>> FreeSWITCH-dev(a)lists.freeswitch.org >>> http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev >>> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev >>> http://www.freeswitch.org >>> >> >> >

2 1

using freeswich as for SIP signalling and programming an external HW.
by Narendra Sirugudi 12 May '11

12 May '11

Hi All, I want to use freeswich for SIP signalling and program an external HW for RTP processing. Does freeswitch provide any mechanism/hooks to program an external HW for RTP processing ? thanks, --kumar

3 2

is there any info abount which git tag/ Freeswitch release is stable?
by Tom Parrott 12 May '11

12 May '11

Hi, I work for a company selling call tracking services, and we have recently put into production a 3 node Freeswitch cluster to use features such as call recording and call announce. We are using the event socket feature to control freeswitch from our custom written service. I have found Freeswitch a pleasure to use, and much easier to get along with than Asterisk. However one thing that has concerned me is the lack of regular point-in-time releases. When putting services in production, we use RPMs to allow us to easily upgrade (and if need be rollback) software versions. The 1.07 version available at http://latest.freeswitch.org seems to be rebuilt each night, does this mean that a 1.07 built yesterday is different from one built today? If I can offer any assistance with the automation of point-in-time tagged releases I would be happy to help, as I feel this would make Freeswitch more attractive to commercial operations as it would easily fit into well defined release processes. Thanks Tom

2 2

FreeSWITCH Conference Call Today
by Michael Collins 11 May '11

11 May '11

Hello gang! We have a light agenda again today, however I still want everyone to call in so that we can discuss a few topics. The agenda is here: http://wiki.freeswitch.org/wiki/FS_weekly_2011_05_11 Be sure to add any items that you want to discuss today. -Michael

1 0

Re: [Freeswitch-dev] Noise heard when a user holds a conference in SRTP mode
by Ujjwal SIngh 10 May '11

10 May '11

Hi Brian, > > > I feel that the key exchange is fine in the way the hold is taking place. > > (FreeSwitch) > A CONF IVR > |ReINVITE F1(w/o SDP)| | > |------------------------------->| | > > | 200 (with SDP) (K1) | | > |<-------------------------------| | > | INVITE(K1) F3 | | > |---------------------------------|----------------------->| > | 200 (K2) F4 | | > > |<-------------------------------|-------------------------| > | ACK F5 | | > |---------------------------------|----------------------->| > | ACK (SDP)(K2) F6 | | > |-------------------------------->| | > | | | > | | | > | | | > | | | > (SRTP key K1) (SRTP key K2) > > As you can see in the above call flow : > > The local key of conference generated in the offer(200 OK) of the ReINVITE > is used as the local key by the endpoint for the INVITE sent to MoH, and the > key generated by the MoH in the answer(200 OK) is used as the key in the > answer of the ReINVITE(ACK). > > I have removed the wireshark trace, the mail wasn't getting through :'( > > > F1 : (ReINVITE w/o SDP) > > F2 : (200 OK with SDP) Key K1 > > F3 : (INVITE with SDP) Key K1 > > F4 : (200 OK with SDP) Key K2 > > F5 : ACK > > F6: (ACK with SDP) Key K2 > Please let me know if I am missing something here. > > Regards, > Ujjwal > > On Thu, Apr 21, 2011 at 7:38 PM, Brian West <brian(a)freeswitch.org> wrote: > >> Yah SRTP might NEVER work with that method of hold because now the MOH >> server and the endpoint have to exchange SRTP keys when you refer to them >> and they probably are NOT doing that right now which is why you get white >> noise. >> >> /b >> >> On Apr 21, 2011, at 6:26 AM, Ujjwal SIngh wrote: >> >> Yes Brian its a white noise, but the hold method works perfectly fine when >> used with RTP, the hold is >> >> implemented according to the draft >> >> http://datatracker.ietf.org/doc/draft-worley-service-example/?include_text=1 >> >> >> Regards, >> Ujjwal >> >> >> >> >> On Thu, Apr 21, 2011 at 3:13 AM, Brian West <brian(a)freeswitch.org> wrote: >> >> I'm going to guess your device method of placing you on hold is wrong. >> And >> >> possibly doesn't encrypt the data or doesn't signal a key change... is it >> >> white noise? >> >> >> /b >> >> >> >> _______________________________________________ >> FreeSWITCH-dev mailing list >> FreeSWITCH-dev(a)lists.freeswitch.org >> http://lists.freeswitch.org/mailman/listinfo/freeswitch-dev >> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-dev >> http://www.freeswitch.org >> >> >

1 0

is there any info abount which git tag/ Freeswitch release is stable?
by Tom Parrott 09 May '11

09 May '11

Hi, I work for a company selling call tracking services, and we have recently put into production a 3 node Freeswitch cluster to use features such as call recording and call announce. We are using the event socket feature to control freeswitch from our custom written service. I have found Freeswitch a pleasure to use, and much easier to get along with than Asterisk. However one thing that has concerned me is the lack of regular point-in-time releases. When putting services in production, we use RPMs to allow us to easily upgrade (and if need be rollback) software versions. The 1.07 version available at http://latest.freeswitch.org seems to be rebuilt each night, does this mean that a 1.07 built yesterday is different from one built today? If I can offer any assistance with the automation of point-in-time tagged releases I would be happy to help, as I feel this would make Freeswitch more attractive to commercial operations as it would easily fit into well defined release processes. Thanks Tom

1 0