Re: Adding parallelism? (was Re: Cost vs. Benefit of Real-Time Applications and Infrastucture Area)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Pekka,

Actually that has been discussed, as have the scope boundaries
at the "top" (apps) and "bottom" (sub-IP). And the diffculty always
is the need for cross-fertilization and cross-area review.

Do you think that applications protocols can be designed with
only a liaison relationship to transport, or transport with
only a liaison relationship to the network layer?

Anyway, thanks for changing the subject line for a new subject.

I think the answer lies in delegation, which is the way management
structures should handle parallelism and overload. But I'm getting
ahead of the PESCI process there.

   Brian

Pekka Nikander wrote:
On Sep 21, 2005, at 7:25, David Kessens wrote:

I would have a lot less trouble with the proposal of adding an area if
we would be able to find another one that could be abolished, or
reorganize ourselves in some way or form that would result in no net
addition of Area Directors.


On Sep 21, 2005, at 21:14, David Kessens wrote:

First of all I don't believe that 15 members in the IESG is manageable
in any way or form. In fact, I believe that 13 members is already not
manageable.


Have we (== the IETF) ever seriously considered splitting this
organisation "horizontally"?  That is, instead of having one IETF
worrying about all from sub-IP issues to (some) applications, have
two ones, one more focusing on issues at the IP layer and directly
above and below it, and the other one on issues "above" that?

Maybe doing that *all* the way, with two weeks of meetings, two IESGs,
two IABs, two SEC and OPS areas, etc?  Of course, that might be very
very bad for the Internet, and would certainly be burdensome for some
regular IETFers, basically doubling their time spend at IETF meetings.

However, <emph> might the benefit be greater than the cost? </emph>

If we think that a full split is unlikely to work out, might there
be other ways of adding management parallelism, preferably without
adding management depth?  For example, how important is it that
documents get reviewed by *all* current areas?  There are certainly
some "cross-layer" areas, such as SEC and OPS, that should worry
about all layers, but aren't the primary purpose of layering to
make sure that people working on the other end the stack can feel
safe to _mostly_ ignore details at the other end; e.g., that apps
people don't need to worry about the details of the routing?

I realise that any "horizontal" splitting is likely to cause
architectural problems in the long run, but in my very humble
opinion we are already now doing pretty badly w.r.t. the overall
evolution of the architecture.  In other words, I don't believe
that the current organisation and division of work is good for
the architecture, but that is a separate issue.  [In this mode of
splitting perhaps the IAB could be split vertically, forming an
Internet Administration/Appointments Board and an Internet
Architecture Board :-)]

Anyway, I am just trying to think outside of the box; this
suggestion should not be taken too seriously.  And perhaps this
all should even be discussed somewhere else and some other time.

My 2 cents.

--Pekka Nikander


_______________________________________________

Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf



_______________________________________________

Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf

[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]