[Serdev] Concept for a SIP cluster implementation

[Jan Janak]

On 27-09-2005 15:16, Klaus Darilion wrote:
> Andreas Granig wrote:
> >Klaus Darilion wrote:
> >
> >>Jan Janak wrote:
> >>
> >>>  - A user agent does not have to send INVITE messages through the same
> >>>    proxy instance as REGISTER messages. The proxy server that received
> >>>    the INVITE request can make sure that the uac is reachable
> >>>    for the duration of the transaction by retransmitting the
> >>>    provisional response periodically (simulates NAT ping). All other
> >>>    transactions finish immediately and do not need this.
> >>
> >>
> >>
> >>What happens in the callee hangs up (send the BYE)? This will be 
> >>loose-routed via the same proxy, which may not have a pinhole to 
> >>traverse the NAT :-(
> >
> >
> >Not if the proxy which receives the first INVITE from the caller sets a 
> >record-route and NAT-pings the caller until the BYE is received. But 
> >this would require some sort of dialog-stateful mechanism.
> 
> So, we just have to make ser call-stateful first :-)
> 
> >I'd say it's much easier to handle if the outbound proxy equals the 
> >registrar from the UACs point of view. So the UAC always sends all 
> >requests to the same edge proxy, which then splits up REGISTERs from 
> >other requests and forwards them to the appropriate server.
> 
> Of course it would be much easier. How will you handle load blalancing 
> between the edge proxies? SRV? If you use SRV, the clients are free send 
> the request to whatever edge proxy. RFC 3263 says to use SRV lookup for 
> each transaction.

  Redirect them to the right proxy, but only when needed.

    Jan.