On Wed, Apr 1, 2009 at 11:27 AM, Scott Lawrence
<[email protected]> wrote:
> On Tue, 2009-03-31 at 21:40 -0400, M. Ranganathan wrote:
>> 2009/3/31 Alfred Campbell <[email protected]>:
>> > I was just look thought the posts and didn’t see one that talked about 
>> > this.
>> > It seemed Tony was seeing some of this I believe.
>> >
>> > Problem:  Some ITSPs use port 5060 and if that site also requires to use
>> > Remote Worker this can be nothing short of challenging to configure.  With
>> > no ability to change the proxy port via the UI how do we think this is to 
>> > be
>> > configured in the field? I believe someone mentioned a DNS option to solve
>> > this however with the challenges we see people having with DNS is this 
>> > going
>> > to be feasible?
>> >
>> > Since bandwidth.com is very popular should we not look to make this easier
>> > to setup?
>> >
>> > Al
>>
>>
>> I think it would make configuration much simpler to be able to run the
>> proxy at a port other than 5060 and make that be configurable via
>> sipxconfig.
>>
>> In addition, It might make sense to add a statless proxy capability to
>> sipxbridge. This way, if sipxbridge can identify requests originating
>> from remote workers it can statelessly proxy these requests to the
>> proxy server ( no relays, SIP Transactions or Dialogs involved - just
>> add a Via header and forward the request. Strip the via header for
>> responses to proxied requests and forward them ). A single public
>> signaling port needs to be configured and exposed through the firewall
>> and that port can be 5060.  There is no need for split DNS.
>>
>> The two approaches are not mutually exclusive.
>>
>> The down side to the stateless proxy is that it winds up being a
>> single point of failure. Our redundancy solution of registering the
>> remote worker phone with each of the redundant proxy servers will not
>> work any more. It also adds signaling overhead.
>>
>> P.S. This just in from Tony: Bandwidth.com apparently does allow you
>> to set the port so you can make it send signaling at any port you
>> want. This temporarily solves our problem. Others may not be so
>> flexible.
>
> I'd say that last bit of news is sufficient to consider this closed for
> now.
>
> We should probably think about the bridge/proxy solution for the long
> run so that we have only one thing running on the outside...
>
>

OK I will open up an issue for consideration in 4.2 so we can revisit it then.

Thanks for responding.

Ranga

>



-- 
M. Ranganathan
_______________________________________________
sipx-dev mailing list
[email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev

Reply via email to