Dear Eric
On Jul 23, 2008, at 12:25 PM, Eric Burger wrote:
Can you enumerate the various options? Thanks.
See the attached email which describes them - here is the list with a
description of each. If you can think up more,
let me know.
Marshall
Option Description Extra Rough Cost / Year
Biggest Drawback IMHO
Slots / year
1 More Friday Sessions 48-72 $ 100,000 to 140,000
Staying Friday night
2 After Dinner Sessions 108 $ 90,000
Staying late at night
3 9 simultaneous sessions 48 $ 140,000
AD availability, collisions
4 4 meetings / year 117 highly profitable
Extra travel, would have to wait till at least 2010
4.a 4 meetings, no Friday 97 profitable
Extra travel, would have to wait till at least 2010
5 Area interim meetings 24-32 neutral or profitable
Extra travel, inefficient for attendees
From an email to this list, July 17, 2008 9:48:22 PM EDT
Dear Russ;
After our discussions on this in San Jose, I spent a little time
thinking of options for extra meeting time. Here are some more
considered thoughts, focusing mostly on costs and meeting logistics,
and intended to engender further discussion. I will be ruthless in
doing back of the envelope cost estimates in the absence of actual
ones. A detailed financial statement for IETF-70 (Vancouver) is
available at
http://iaoc.ietf.org/documents/Vancouver_Financial_Statement_Preliminary.pdf
and I use that for price estimates.
It seems to me that the IETF is bursting at the seams. Certainly both
the number of interim meetings seems to be increasing (are statistics
on the number and attendance of interim meetings available anywhere?),
so it is reasonable to try and find means that will provide more
meeting slots.
Possible solutions include
1 - Add meetings to Friday (the proposed experiment). We could add 3,
but the experiment is for 2. If we kept the rest of the meeting the
same, then the 2 session option would add 16 sessions per meeting, or
48 per year. If we add 3 sessions, there would be 24 per meeting, or
72 per year. There would definitely be extra expenses for these
options :
- Currently, the IETF is torn down starting at 1130 AM on Friday. This
obviously would not be possible with this option, and there would
almost certainly be an extra Verilan (Connectivity and NOC) and maybe
a AMS (Secretariat) expense, which presumably Ray could estimate or
get quotes for. For Vancouver, Verilan was $ 115,000, so an extra half
day would be about $ 25,000.
- There would be an extra hotel charge, particularly if the Hotel is
unable to use our space for Friday evening parties. This might show up
(especially in the US) and an increased food and beverage cost even if
we don't provide refreshments, and Ray should see if that is the case
for Minneapolis. Likewise, there might be $ 5000 in Audio Visual costs.
- Since there are existing Hotel contracts for a number of future
meetings, these would have to be renegotiated, and it is likely that
some of these hotels will now be booked for Friday afternoon (or will
have Friday evening parties that will need the setup time), so this
option may not be possible for all of our future meetings.
- If the Friday meeting goes late enough that people have to get a
Friday night hotel room, then both participants and the IETF will have
a fairly considerable extra expense.
If we just assume that charges are linear with meeting time, this
option would entail something between $ 35,000 and
$ 48,000 (based on Vancouver), or $ 105,000 to $ 140,000 / year or
about $ 40 added to the meeting fee.
I would urge Ray to prepare a proper budget with actual quotes for
this option.
2 - Add evening sessions to the existing meetings, presumably after a
dinner break. This would add 8 x 4 = 36 slots per meeting, or 108 per
year. It would probably not cost the IETF much (it is not clear to me
if Verilan would charge more for the longer hours; the total charge
for this in Vancouver was $ 115,000; the extra cost would ~ $ 30,000
($90,000 / year) per meeting assuming a linear charge for the extra
time. I would expect the actual extra charge to be less; this could be
determined by Ray while he was doing the estimates for # 1. I know
that people did not like the meetings after dinner, so this would
probably be resisted, but it would add more slots and probably at a
lower price. I don't know if there would be extra hotel charges, but I
suspect not.
3 - Go from 8 simultaneous sessions to 9. This would add 16 sessions
per meeting, or 48 per year with the Dublin scheduling. This option
would of course make collisions worse, and might be limited by AD
availability. It would also require more hotel rooms, and some
upcoming hotels might not be able to support it. Basically all of the
costs (hotel, Verilan, etc.) would increase. Again assuming a linear
model for costs, the extra cost would be about $ 48,000 per meeting,
or $ 144,000 per year, or an extra $ 50 or so to the registration fee.
4 - Add a fourth meeting. This would provide ~ 117 slots per year.
This would actually considerably increase the IETF's revenue _if the
attendance stays more or less the same_. As you know, the IETF
meeting income is used to pay for the other IETF activities. IETF-70
provided a profit of $769,534, so this solution would probably
increase the IETF's yearly net income by $ 500,000 or more. I have a
feeling that many people would resist this, and of course it would
probably not be possible to schedule a fourth meeting before 2010, if
then.
5 - Having some sort of structured interim meeting or meetings during
the year. With 8 Areas, there could be 3 or 4 mini IETF's during the
year. If these area-interims lasted 2.5 days each, with no
simultaneous sessions, each might add 8 slots, or say 24 to 32 per
year. I would assume that these meetings would be run to at least
recoup costs if not provide additional income for the IETF. These
meetings, being small, would be fairly easy to book, and probably
could start more or less immediately.
Note that, while no one might feel that they had to attend every area-
interim, many might feel that had to attend more than one, so in some
ways this would cause more travel for attendees than option 4, for
many fewer additional slots. This option has, however, the great
advantage that WG could get a lot more focused time to work on issues.
So, here is a table on the options
Option Slots / year Rough Cost / Year Biggest Drawback IMHO
1 48-72 $ 100,000 to 140,000 Staying Friday night
2 108 $ 90,000 Staying late at night
3 48 $ 140,000 AD availability,
collisions
4 117 highly profitable Extra travel, would
have to wait till at least 2010
5 24-32 neutral or profit Extra travel,
inefficient for attendees
One danger for some of these options is that there would probably be
more pressure to structure the meetings so
that not everyone has to come to every day of the meeting, which could
have unpredictable results.
So, in conclusion, I would
- support the experiment
- request better cost estimates for the experiment
- suggest that in the actual operations the experiment would need to
be expanded to give most WG even one extra meeting per year
- suggest that combinations of options 1,2,3 and 5 will need to be
explored unless we are willing to move to option 4.
I hope you find this reasoning useful.
Regards
Marshall
On Jul 23, 2008, at 12:14 PM, Marshall Eubanks wrote:
Dear Eric;
On Jul 22, 2008, at 4:25 PM, Eric Burger wrote:
I will take a swag...
On the one hand, having had (and, believe it or not, this time by
request) work group meetings on Friday morning, the Friday slot is
an excellent filter to ensure that only the die-hard participants
and those people paying their own way who need a Saturday Night
Stay(tm) attend.
On the other hand, this guarantees more than a week away from
home. If you have no family or life or want to stay away from
said family or life, this would be a good thing. However, for the
rest of us, this is not an attractive option.
Since Pandora's Box is now formally open, how about this: since
the Internet Area has a lion's share of slots, and RAI draws huge
crowds, what if we went summer / winter olympics and (gasp!) add a
fourth annual meeting, but not have all areas meet at all meetings.
This is my option 4. I think that a variant, call it 4.a, might
also be useful :
Have 4 meetings per year, but close the meeting down Thursday
evening. This would save a little money per meeting, raise
more money for the IETF (by having 4 meetings) and would mean that
for most of us, the IETF would only eat 1 weekend, not 2.
Here are the stats. I don't think that adding even a 4 day meeting
could be scheduled before 2010, if then.
Option Extra Rough Cost / Year Biggest Drawback IMHO
Slots / year
1 48-72 $ 100,000 to 140,000 Staying Friday night
2 108 $ 90,000 Staying late at night
3 48 $ 140,000 AD availability,
collisions
4 117 highly profitable Extra travel, would
have to wait till at least 2010
4.a 97 profitable Extra travel, would
have to wait till at least 2010
5 24-32 neutral or profit Extra travel,
inefficient for attendees
Regards
Marshall
Yes, for me that just means four meetings to go to instead of
three. However, one would hope we could get the work done Monday -
Friday, instead of Saturday - Saturday, which we really are
talking about here. Moreover, there really are people that will
never venture outside their area, no matter how much we tell them
it is important for them, and the IETF as a whole, to do it. Thus,
rather than fighting physics, let us embrace gravity and make for
a more sensible schedule all around.
On Jul 17, 2008, at 5:33 PM, IETF Chair wrote:
The IESG is considering an experiment for IETF 73 in Minneapolis,
and
we would like community comments before we proceed. Face-to-face
meeting time is very precious, especially with about 120 IETF WGs
competing for meeting slots. Several WGs are not able to get as
much
meeting time as they need to progress their work. As an
experiment,
we are considering adding two Friday afternoon one-hour meeting
slots.
The proposed Friday schedule would be:
0900-1130 Morning Session I
1130-1300 Break
1300-1400 Afternoon Session I
1415-1515 Afternoon Session II
Please share your thoughts about this proposed experiment. The
proposed experiment will be discussed on the IETF Didcussion mail
list (ietf@xxxxxxxx).
_______________________________________________
IETF-Announce mailing list
IETF-Announce@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf-announce
_______________________________________________
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf
_______________________________________________
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf