RM impl slides and algorithm

[jacques durand <jacques@savvion.com>]

Hi all:

Here is a more complete version of my RM implementation slides,
where I detail further a new "Un-Ordered Sequence Number Method" that
we started to discuss at the f2f this week, and
that seems suitable for reliable concurrent message sending.
I have also addressed  what happens when the sequence counter wraps
around
(a condition that usually should not occur before sending many millions
messages,
so I assume given the tight schedule in Tokyo we won't have a chance to
demo this!).

Note the copyright notice: my main intent in making it explicit on this
detailed
material (and in publishing the latter) is in fact to keep it public.
(until if/when it becomes part of some more formal ebXML/Oasis doc
material,
e.g. non-normative appendix to RM spec.)
It should be freely usable for implemention, nobody should try to patent
it...
(now, as I am not an expert in messaging, could very well be that this
is nothing new!!)
I see it as a way to efficiently implement (and validate) the current RM
spec.

Please review and feel free to comment/implement/test.

Regards,

Jacques Durand
Savvion

RMslides.ppt