Frank W. Miller | 1 Apr 05:55 2006

Available services supporting NAPTR queries


Greetings,

Does anyone know of any Service Providers or other domains that support
NAPTR queries?  I need to bounce some NAPTR queries off a domain for
testing.  One with multiple defined services such that you're forced to
parse multiple lines to get what you want would be great.

Also, I've run SRV queries at fwd.pulver.com and they seem to work but
they only specify a single server address.  Anyone know of a service
that includes multiple server addresses in an SRV response?

Thanks,
FM

Joegen E. Baclor | 1 Apr 06:18 2006

Re: Question regarding Record-Route

Hi Manju,

Thanks, I totally agree with your explanation.  I think it is obvious 
that my real intention is to make sure that the proxy receives the BYE 
which brings me to the next (dirty) alternative.  Modify the Contact 
header in the INVITE and replace it with the proxy's address before it 
goes down the path.  If the proxy can cache the original Contact header 
and use it to route suceeding messages, would that be an allowable option?

Joegen

Manjunath Warad wrote:

>Hi Joegen,
>	    
>	If the UAS does not comply to the specification behavior and
>dispatches the 200 response without the RR(Record-Route), then the UAC will
>use the remote target directly to send the message, bypassing all the
>Record-Routing Proxies. Take it another way, a misbehaving UAC can also
>choose to ignore the RouteSet even if present in the 200 response and stll
>try to send the message directly to the remote target. So the question
>becomes whether the signallling should go through ?
>
>	The standards do not require the proxy to manipulate the Record
>Route headers in the response. If a network is configured in such a way that
>UAC and UAS cannot directly communicate with each other, but able to do so
>only if they go through these Record-Routing proxies (that's the reason that
>they would record-route in the first place), then there should be no problem
>and the call will always fail without creating any billing issues also.
>
(Continue reading)

M. Rangnathan | 1 Apr 15:29 2006

SDP Offer/Answer and SSRC.

Hello!

I am building an RTP point to multipoint multiplexer. For this, when a 
packet comes in, I need to send it in to all receivers excepting the one 
which sent me the packet. It would be conveniant if I knew the ssrc of 
all receivers in order to do this but how do I know this before the 
packet is sent?

Hence, I am looking for some information which allows me to extract the 
SSRC of the receiver of an RTP session from the SDP answer. Is such 
information present in the SDP answer and where?   I sifted through the 
SDP offer/answer RFC but I dont seem to find any information about this 
issue.

Another way I can do this is to rely on IP address and port and make 
assumptions such apriori knowledge of IP and port where the sending 
socket will be bound ( this way I can look it up in the incoming RTP 
packet and avoid routing the packet back to the sender ). However, I 
dont find any information in the RTP RFC that obliges the sender to bind 
to any speicific IP address and port (only the receiver needs to bind to 
the IP and port in the SDP answer and not the sender). Please correct me 
if I am wrong in this assumption.

   
Thanks.

Ranga

--

-- 
M. Ranganathan 
(Continue reading)

Dale R. Worley | 1 Apr 21:35 2006

Re: Question regarding Record-Route

On Sat, 2006-04-01 at 12:18 +0800, Joegen E. Baclor wrote:
> Thanks, I totally agree with your explanation.  I think it is obvious 
> that my real intention is to make sure that the proxy receives the BYE 
> which brings me to the next (dirty) alternative.  Modify the Contact 
> header in the INVITE and replace it with the proxy's address before it 
> goes down the path.  If the proxy can cache the original Contact header 
> and use it to route suceeding messages, would that be an allowable option?

What are you trying to accomplish?

If the UAs obey the standard, the proxy only needs to add a Record-Route
for itself.

Are you worried about interoperating with UAs that do not obey the
standard?

Dale

--- 
interop.pingtel.com -- the public SIP phone interoperability test server

Dale R. Worley | 1 Apr 21:42 2006

Re: Available services supporting NAPTR queries

On Fri, 2006-03-31 at 22:55 -0500, Frank W. Miller wrote:
> Does anyone know of any Service Providers or other domains that support
> NAPTR queries?  I need to bounce some NAPTR queries off a domain for
> testing.  One with multiple defined services such that you're forced to
> parse multiple lines to get what you want would be great.

Why don't you set up your own DNS server, so you can control exactly
what sort of horrors DNS queries return?

Dale

--- 
interop.pingtel.com -- the public SIP phone interoperability test server

Joegen E. Baclor | 2 Apr 03:57 2006

Re: Question regarding Record-Route

Hi Dale,

Yes, I am trying to overcome the fact that many UA's do not honor record 
route when they are present.   This is a big problem for proxies 
responsible for call accounting.   I know one solution would be to go 
full B2BUA.  However, I want to remain as none-intrusive as possible 
with the headers so as not to open new cans of worms with regard to 
interoperability between legs.   All I want is to be sure that I get a 
BYE or any other means for the proxy to know that the call is still 
persisting, period.   Any ideas how this is accomplished in the real world?

Thanks for the attention!

Joegen

Dale R. Worley wrote:

>On Sat, 2006-04-01 at 12:18 +0800, Joegen E. Baclor wrote:
>  
>
>>Thanks, I totally agree with your explanation.  I think it is obvious 
>>that my real intention is to make sure that the proxy receives the BYE 
>>which brings me to the next (dirty) alternative.  Modify the Contact 
>>header in the INVITE and replace it with the proxy's address before it 
>>goes down the path.  If the proxy can cache the original Contact header 
>>and use it to route suceeding messages, would that be an allowable option?
>>    
>>
>
>What are you trying to accomplish?
(Continue reading)

Frank W. Miller | 2 Apr 06:53 2006

Re: Available services supporting NAPTR queries


Well, the purpose of my question was twofold.

First, if such servers are already setup, its always good to test
against something you don't control.  That way, you're not biasing the
testing in any way.  This is the main reason why I've never setup a STUN
server in my own environment.

Second, if the answer comes back that nobody is setting up the service,
there's not much point in implementing it is there?

Besides, if theres something out there already, I'd like to avoid the
"horror" of setting bind or something else if I can.

FM

On Sat, 2006-04-01 at 14:42 -0500, Dale R. Worley wrote:
> On Fri, 2006-03-31 at 22:55 -0500, Frank W. Miller wrote:
> > Does anyone know of any Service Providers or other domains that support
> > NAPTR queries?  I need to bounce some NAPTR queries off a domain for
> > testing.  One with multiple defined services such that you're forced to
> > parse multiple lines to get what you want would be great.
> 
> Why don't you set up your own DNS server, so you can control exactly
> what sort of horrors DNS queries return?
> 
> Dale
> 
> --- 
> interop.pingtel.com -- the public SIP phone interoperability test server
(Continue reading)

Niranjan Gopalakrishnan | 1 Apr 02:15 2006

Multi homed Proxy

Im working with an implementation of a Multi homed proxy - uses 2
interfaces, I presume one external and one internal. 
On receiving a request on one interface, it forwards it to itself on the
destination interface, eventually adding itself twice in the
Record-Route header (with r2 parameter) before forwarding the request.
Response is processed similarly.

Why is such a behaviour required? If this is to seperate the netowrk
topology (external, internal) the same can be achieved by an IP gateway.

I am sure there is only one instance of the Proxy running on the host.

This behaviour is not affecting our functionality. But I need to
understand the reason behind it.

Any pointers appreciated.

Thanks.
Niranjan Gopalakrishnan
Senior Engineer, Call Control, Veraz Networks.
niranjan <at> veraznet.com

Kedar Karmarkar | 2 Apr 23:21 2006
Picon

Re: SDP Offer/Answer and SSRC.

 Hello,

Could you please explain what kind of call control you are using to setup
this conference? Assuming you are using SIP, you will have a port and IP
address on the input side on your MCU, and you will exchange this SDP
information with other participants, wherein you will receive their SDPs,
and hence the forwarding is going to be mostly mechanical. Why do you need
SSRC information?

On 4/1/06, M. Rangnathan <mranga <at> nist.gov> wrote:
>
> Hello!
>
> I am building an RTP point to multipoint multiplexer. For this, when a
> packet comes in, I need to send it in to all receivers excepting the one
> which sent me the packet. It would be conveniant if I knew the ssrc of
> all receivers in order to do this but how do I know this before the
> packet is sent?
>
> Hence, I am looking for some information which allows me to extract the
> SSRC of the receiver of an RTP session from the SDP answer. Is such
> information present in the SDP answer and where?   I sifted through the
> SDP offer/answer RFC but I dont seem to find any information about this
> issue.
>
>
> Another way I can do this is to rely on IP address and port and make
> assumptions such apriori knowledge of IP and port where the sending
> socket will be bound ( this way I can look it up in the incoming RTP
> packet and avoid routing the packet back to the sender ). However, I
(Continue reading)

Mudigonda, Saisrikiran | 3 Apr 08:06 2006

Re: Multi homed Proxy

Niranjan,

Multi-homing and multi-streaming are generally used when your proxy acts
as a B2BUA. These are used because B2BUA is a single point of failure
and hence High Availability needs to be added for robustness. 

My guess is your proxy is acting like a B2BUA. Can you check your proxy
behavior?

Thanks,
Sai.

-----Original Message-----
From: sip-implementors-bounces <at> cs.columbia.edu
[mailto:sip-implementors-bounces <at> cs.columbia.edu] On Behalf Of Niranjan
Gopalakrishnan
Sent: Saturday, April 01, 2006 5:45 AM
To: sip-implementors <at> cs.columbia.edu
Subject: [Sip-implementors] Multi homed Proxy

Im working with an implementation of a Multi homed proxy - uses 2
interfaces, I presume one external and one internal. 
On receiving a request on one interface, it forwards it to itself on the
destination interface, eventually adding itself twice in the
Record-Route header (with r2 parameter) before forwarding the request.
Response is processed similarly.

Why is such a behaviour required? If this is to seperate the netowrk
topology (external, internal) the same can be achieved by an IP gateway.

(Continue reading)


Gmane