Re: Root level prodir combined results

[Colin Robbins <[e-mail]>]

   >why not change the title:
   >  probing the status of R&D networks
   >that would be better suited
   >
   >-- PAP


There are two motives behind probing.  The first is to identify DSAs
that are not available due to software failure, so that the DSA
manager can be prompted to take some corrective action.  The second is
to get a general picture of how accessible the overall DIT is.

I think with a small change in the way results are presented, the
current critisims could be addressed, and perhaps we will move a step
closer to addressing the two issues.

The first point is done reasonably well by the bottom half of the
current probe results.  In future, the report we mail each week will
only show DSAs available less than 75% of the time.  Last week this
would have meant the following:

 62.77 UNINETT, Boa Constrictor                                     411
 59.51 University of Zagreb - Faculty of Electrical Engineering, 
			Roufous Motmot				    410
 56.48 U.S. DMD, Fruit Bat                                          409
 55.47 NEXOR Ltd, Vampire Bat                                       411
 46.34 U.S. DMD, Alpaca                                             410
 31.05 North Atlantic Treaty Organization, Violaceous Trogon        409
 22.63 IIT Delhi, IN, Pelican                                       411
  0.00 Master DSA of KOREA, Darter                                  409
  0.00 Bell-Northern Research, Pangolin                             409

(My excuse as "Vampire Bat" manager: the NEXOR IP link was down for a
long period last week!  This is what dragged the rest of the results
down as well) 

The second point is harder to deal with.  The probe results have some
useful data in there somewhere, but it is hard to interpret.  A better
way would be to plot the information of a graph.  In the following
message will be a example of such a graph using last weeks data
(showing the main problem was indeed NEXOR IP link).  I can't promise
we'll be able to do this every week, it depends on how easy it is to
automate the process, but if there is sufficient demand...

However, this graph is still only looking at DSA availability, it
really needs to be looked at in conjunction with something like Paul
Barkers results,  which show data availability (in this case as seen
from a particular DUA).


Colin


PS  If anybody wants to see the full set of merged results let me know
(please don't copy the list) and we will either set up a
"probe-result" mail list or put the results in the WWW.