H.323 Implementers Guide Edits

Dave Walker dave_walker at Mitel.COM
Tue Apr 20 15:28:54 EDT 1999

As far as I'm concerned, a lightweight RRQ from an unregistered
endpoint is about as valid as any other unexpected message from
unregistered endpoints, i.e. invalid, and I have a strong urge
to ignore them.

OTOH, if the Gatekeeper is interested in responding to lightweight
RRQs subsequent to expiration of a registration, it's free to maintain
the registered RAS address and endpoint id for as long as it wants.
In addition, the politeness of such a Gatekeeper probably means that
it should send a URQ to the endpoint prior to deletion of the
registration information, so the situation should never arise.

Re-use of an endpoint id following deletion of a registration could
cause problems where a Gatekeeper is indexing lightweight RRQ based
on endpoint id.  If a lightweight RRQ arrives from an expired endpoint
after its endpoint id has been assigned to another endpoint, as
currently defined the RCF (or RRJ) would end up being sent to the
currently registered endpoint.  This is a legitmate problem, although
it's unlikely to arise and easy to design around without necessitating
the addition of a response address to the RRQ.


Dave Walker
Mitel Corporation
Ontario, CANADA

Espen Skjæran wrote:
> Jim,
> Concerning the use of lightweight RRQ:
> A lightweight RRQ consists of (ref h225.0, 7.9.1 : "keepAlive, endpointIdentifier,
> gatekeeperIdentifier, tokens, and timeToLive")
> A gatekeeper receiving a lightweight RRQ in cases the registration has aged (and is
> removed) has no information of where to reply to this RRQ. Using the gk discovery port
> and RRQ sender's IP address is not a good option.
> Can we add rasAddress to this list in section 7.9.1 of H.225.0 ?
> Espen

More information about the sg16-avd mailing list