[Asterisk-Users] Asterisk SIP calls failing - not a proxy? What of RTP codec transcoding?

[Ian Blenke]

Fixed it. Looking back a few emails, someone mentioned SIP natting.. so 
this appears to have fixed my problem:

	sip.conf:

	[general]
	register => XXXXX:password at fwd.pulver.com/1000
	
	;; Free World Dialup Proxy
	[fwd.pulver.com]
	type=friend
	host=fwd.pulver.com
	fromuser=XXXXX
	fromdomain=fwd.pulver.com
	secret=password
	username=XXXXX
	context=incoming
	nat=yes
	reinvite=no
	canreinvite=no

and to accept incoming calls correctly?:

	extensions.conf:

	[incoming]
	exten => s,1,Dial(SIP/phone1&SIP/phone2,20,tr)
	exten => s,2,VoiceMail,u1000
	exten => s,102,VoiceMail,b1000

Anyway, the above seems to work for me.

- Ian

Ian Blenke wrote:
> I have an Asterisk 0.4.0 install working with two grandstream budgetone 
> 100 phones, gnophone, and kphone. This is a private network segment 
> (172.17.x.x), with the PBX configured on my outbound firewall which has 
> a public address (66.x.x.x).
> 
> - I can make calls between phones - all extensions are working.
> - I can make IAX calls to IAXTEL. No problems (apparently gsm only)
> - I can call SIP phone numbers
>   - The called party can hear me.
>   - I cannot hear them.
> 
> After looking at the SIP handshaking, it is apparent that Asterisk is 
> giving out my Grandstream's private IP address (172.17.x.128) to the 
> called party in the INVITE. This is bad.
> 
> The documentation suggests that Asterisk terminates all RTP streams and 
> does codec transcoding to make negotiated calls to external SIP 
> endpoints - very proxy like (the behavior I'm looking for).
> 
> How do I configure Asterisk to "hide" origionating SIP phone addresses, 
> masquerading as itself instead?
> 
> I guess my only option is going to be an Asterisk install on the public 
> Internet with *no* private connection and some kind of SIP proxy on my 
> firewall firewall (behind which the phones will sit). All phones will 
> need to register with the Asterisk PBX through the outbound proxied 
> connections.
> 


-- 
- Ian C. Blenke <icblenke at nks.net>
(This message bound by the following:
http://www.nks.net/email_disclaimer.html)