At 12:14 PM 9/3/2003, Gary Brewer wrote:
>Thanks Jan,
>
>I should have made myself a little bit clearer; I am using the RTC 1.2 API
>(http://msdn.microsoft.com/library/default.asp?url=/library/en-us/rtcclnt/rt
>c/real_time_communications_rtc_client_start_page.asp)
>
>What I find the most interesting is -
>http://msdn.microsoft.com/library/default.asp?url=/library/en-us/rtcclnt/rtc
>/traversal_of_upnp_enabled_nats.asp specifically the last sentence of the
>third paragraph. Why I can't provide it my external IP address and port on
>the NAT is annoying! I wonder if it would be possible to modify the REGISTER
>message on the way to the SIP server and change the IP:Port to my external
>NAT address mapping - or get NATHelper to do this for me.
That's exactly what nathelper does with fix_contacts (or something like that).
You need to make sure that you apply it to REGISTER of UACs, who receive
others' requests at the same address from which they send theirs.Otherwise, the external address
taken from singaling's transport is useless.
-Jiri
At 10:55 AM 9/3/2003, Gary Brewer wrote:
>Hi,
>
>I have come across a similar problem. I want to use the A/V facilities of
>Windows Messenger if one or both of my clients are behind a NAT. I realise
>that it seems to be impossible to get this to work if the NAT is symmetric.
>(See: RFC3489 "Applicability Statement")
>
>If only one of my clients is behind a NAT then it would seem I would have to
>communicate my NATs external address and port mapping to the non-NAT'd
>client (possibly with the help of STUN) in my SIP Invite SDP message. I
>would also have to setup UDP mappings for SIP, RTP/RTCP Audio Video on my
>NAT. Are my A/V port mappings also included in the Invite SDP message?
If your client is STUN-enabled, then all ports are patched with their
public equivalents. You don't need to set anything on your NAT.
One-NAT-only is not sufficient for success though -- you still may run
into troubles with symmetric NATs.
>If both clients are NAT'd then what is the approach?
That alone can still work -- it depends on the type of NAT.
>I don't see how I
>register with the SIP server using an external NAT address (my guess is this
>is what I would have to use if I wanted anyone on the other side of the NAT
>to be able to see me).
option 1) have the phone detected the external address and registered with it;
that's what STUN does (e.g., granstream)
option 2) have the server used client's external transport address and ignore
private address inside SIP messages; that's what SER nathelper module
does
>MSFT have seemed to got around this problem by
>recommended everyone to use uPnP enabled NATs, which will automatically bind
>to an external address on the NAT and, I assume, use this when they register
>with the SIP server.
>
>RTPProxy is here https://demo.portaone.com/~sobomax/PortaSIP/ how does
>RTPProxy help in the NAT situation, does it at all?
It does. It uses brute-force: all media hit the proxy, which behaves symmetricaly,
i.e., ignores SDP and sends media to where reverse streams come from. It should
work fairly well with most use cases, the penalty is the bandwidth.
-Jiri
There was a similar post on Aug 13th. I'm having the same problem [http://lists.iptel.org/pipermail/serusers/2003-August/002259.html] and I've already tried the suggested solutions.
Basically, on the user login or the admin login of serweb, when I type a username and password and press login or <enter> the page blanks, reloads to the login page, and does nothing. I get no errors of any kind.
I have register_globals=On in php.ini.
I'm relatively sure my php.ini is in the correct place [i.e. is being used by the server]
How would I check this?
My subscriber table in mysql has the correct domain in the domain column for all my users, including admin. I believe that I've made all the changes in Dan Austins HOWTO for serweb/php configuration.
RedHat 9.0
kernel 2.4.20-20smp
ser 0.8.11 (i386/linux)
main.c, v 1.162.2.5
PHP 4.3.1
Apache 2.0.40
Any ideas?
Thanks,
G.
---------------------------------
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
Hi,
Just to add on my previous posting, where the snom phone didn't have
an RTP stream toward the other end, after doing some more tests
I found that by chaniging who initiates the call I still get a one-way RTP
stream.
What can trigger such a behaviour where a device won't initiate an RTP
stream after receiving an SDP.
Thanks.
Samy.
I am trying to get my setup working with radius. I have the lastest freeradius, radiusclient, and ser.
The test programs that come with radiusclient work fine, so I've determined that the problem now comes down to my ser setup.
I've attached my ser.cfg as well as the output of ser debug.
Any ideas?
<<attempt>> <<ser.cfg>>
Regards,
Joseph Rork
Ford Motor Company
Real-Time Collaborative Applications, SIE
Phone: 313-594-6672
Email: jrork(a)ford.com
"There are 10 types of people in the world: Those who understand binary, and those who don't."
Hello all, I want to make a forward to another uri when something happens.. I saw an script example at the ser site... and I found the script I need.. but I don't know where I have to write the script...? on a separate file... on the config file...? where...
Thanx in advance
Ivan
Hi,
is possible that MSN Messneger limit my online user? I can see only 2 user
online.
thanks. Andera
----------------------------------
Legge di Clarke sulle idee rivoluzionarie
Ogni idea rivoluzionaria - in campo scientifico, artistico o altro provoca
tre stadi di reazione, riassumibili nelle seguenti frasi:
1. "E' impossibile; non farmi perdere tempo".
2. "E' possibile, ma non val la pena di farlo".
3. "L'ho sempre detto, io, che era un'ottima idea".
Hi,
I don't know if this is the right place to post this, but I noticed a strange
behaviour with the snom 100 with ser 0.8.11 and a grandstream.
The snom is behind NAT, but ser is on a public address, the grandstream is on another network also behind nat.
The nat is sip aware (cisco soho91).
When I place a call from the snom to another IP phone, grandstream, the call doesn't always complete successfully: in 50% of cases after the other end answers the call, I can hear the other party, but they cannot hear me.
When I look into ethereal, I find the regular SIP INVITE and rining messages, then I see a
bunch of UDP packets (the RTP session) from the other end, the grandstream, but I don't see a stream of RTP from the snom to the other end.
Now what is strange here is that sometimes everything works fine, and we can both hear each others,
and I do then see the RTP stream in both directions.
Both devices share the same codec, and the SDP info is ok, the ports numbers are valid.
I don't think this is ser related, but I wonder if somebody has experienced such a behaviour.
Thanks.
Samy.
At 12:02 AM 9/4/2003, Guilherme Dal Pizzol wrote:
>ello,
>
>I'm trying to generate call accounting in parallel and sequential forking,
>but I think I'm missing something.
>
>My ser.cfg is attached.
>
>When I perform a call (parallel fork / sequential fork) I'm getting only
>one miss report, for example:
>
>1 - SER receives a INVITE
>2 - INVITE is sent to A and B (parallel)
>3 - A fails with 503
>4 - B fails with 408 (TM generates 408)
>
>5 - SER logs only 408!!! How about 503? Is there any way to account both
>misses?
That's correct -- there is only one final result for a proxied transaction
(no matter how many branches you fork) and that's the value forwarded upstream.
Per 3261, that's module the lowest error code.
-jiri
