< Previous by Date Date Index Next by Date >
< Previous in Thread Thread Index Next in Thread >

Re: [reSIProcate] Getting started


So what you're describing will take something that is fundamentally different from the stack as it exists. You could build something that did what I think you're describing reusing some of the stack's components (parser, scanner, some of the transport code). But it would be a different thing when you were done and would become a separate project as opposed to
something that tried to live in the same library as the stack.

RjS

On Mar 5, 2007, at 3:55 PM, Steve Hanka wrote:

Hi,

The purpose of our stateless proxy is to route calls with load balancing
within a controlled cluster of servers and to ensure high-availability
across a cluster of proxies without requiring the exchange of call state
information between them.

Our application won't use DNS (we have our own routing database) and we won't be maintaining transaction state. We have internal mechanisms for
ensuring that we don't wind up with inadvertent forked calls; if a
repeated INVITE, for example, is received on a different proxy our
mechanisms ensure that it is sent to the same UAS as the original
INVITE.

If repro is dialog stateless and not transaction stateless it doesn't
buy us anything.

Best regards,

Steve Hanka


-----Original Message-----
From: Byron Campen [mailto:bcampen@xxxxxxxxxxxx]
Sent: Monday, March 05, 2007 2:47 PM
To: Jason Fischl
Cc: Steve Hanka; resiprocate-devel@xxxxxxxxxxxxxxxxxxxx
Subject: Re: [reSIProcate] Getting started

        Just to be absolutely clear, when Jason says "stateless proxy
based
on resiprocate called repro", he means _dialog_-stateless, and not
transaction-stateless.

Best regards,
Byron Campen

Here is a snippet of the discussion. You can find more of it using
google. I searched on:
site:list.sipfoundry.org resiprocate stateless proxy

http://list.sipfoundry.org/archive/resiprocate-devel/msg00214.html

The changes required in resiprocate would be substantial. I think that
as long as the api didn't require any major changes, the community
would be supportive.

Are you sure that you need a stateless proxy? Why not use the
existing, stateless proxy based on resiprocate called repro. It should
do everything that you need and is high performance.

It would be helpful to understand what your requirements are that
motivate using a stateless proxy.

Jason



On 3/5/07, Steve Hanka <steve.hanka@xxxxxxx> wrote:
Regarding the flaws in stateless proxies that break the protocol:

Does anyone know, in general, what these flaws are?  The RFCs
specifically mention support for stateless proxies and I am
curious as
to what the problems are.

Would anyone object if I investigated modifying the library to
support
them again?

Best regards,

Steve Hanka


-----Original Message-----
From: jason.fischl@xxxxxxxxx [mailto:jason.fischl@xxxxxxxxx] On
Behalf
Of Jason Fischl
Sent: Monday, March 05, 2007 1:56 PM
To: Byron Campen
Cc: Steve Hanka; resiprocate-devel@xxxxxxxxxxxxxxxxxxxx
Subject: Re: [reSIProcate] Getting started

The original reason that it was deprecated was for the reasons you
listed. One of the consequences of deprecating it is that we stripped
out the support for stateless proxies from the transaction layer of
resip. As a result, it is no longer possible to implement a stateless
proxy in resip without substantial changes to the library.

Jason


On 3/5/07, Byron Campen <bcampen@xxxxxxxxxxxx> wrote:

 As far as I know, the stuff in deprecated is the only stateless-
proxy
code
we have. Anyone remember why specifically this was deprecated? I am
under
the impression it was because stateless proxies in general had
flaws
that
broke the protocol in some situations, and not any specific flaw in
the
code, but I could be wrong.

Best regards,
Byron Campen





Hi,



I am just getting started with reciprocate and am trying to get my
mind
around the system.  For the last week I have been reading the
source
code,
looking at the test code and the API documentation, and so
forth.  My
goal
is to build a stateless proxy/message-relay server with some
special
routing
and message filtering capability for use in my employer's
intelligent
network.



I would like to find some sample code of a simple agent and a
simple
proxy.
I note that these pages are TBD in the documentation pages and that
the
stateless proxy code in the source distribution has been
deprecated.
If
somebody will point me to such sample code, I will gladly write the
documentation pages once I figure out how the code works.



Any suggestions?



Best Regards,



Steve Hanka

Steve.Hanka@xxxxxxx

shanka@xxxxxxxx



 ________________________________


 This message (including any attachments) may
contain confidential information intended
for a specific individual(s) and purpose,
and is protected by law. If you are not the
intended recipient, you should delete this
message.

 Any disclosure, copying, or distribution of this
message, or the taking of any action based
on it, is strictly prohibited.
_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@xxxxxxxxxxxxxxxxxxxx
https://list.resiprocate.org/mailman/listinfo/resiprocate-devel

_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@xxxxxxxxxxxxxxxxxxxx
https://list.resiprocate.org/mailman/listinfo/resiprocate-devel







This message (including any attachments) may contain confidential
information intended for a specific individual(s) and purpose, and
is protected by law. If you are not the intended recipient, you
should delete this message.

Any disclosure, copying, or distribution of this message, or the
taking of any action based on it, is strictly prohibited.







This message (including any attachments) may contain confidential information intended for a specific individual(s) and purpose, and is protected by law. If you are not the intended recipient, you should delete this message.

Any disclosure, copying, or distribution of this message, or the taking of any action based on it, is strictly prohibited.