[CUWiN] HSLS Behavior Question

[Sanjai Narain]

Hi David: In view of your note, it looks like the behavior from HSLS we 
see below is incorrect. Can you please let us know what additional tests 
we could perform to confirm or deny this hypothesis. Can you point us to 
the specific section/files of the code that we could look at? Thanks. -- 
Sanjai

In our testbed that emulates mobility and losses:
    * There are four nodes, 6, 7, 10, 66.
    * Between times 292 and 598, 6 continues to advertise 10, 66, 7 as 
its neighbors in Hello messages
    * Yet 10 does not advertise 6 in its Hello messages.

Since 6 is advertising 10 as its neighbor, it has heard 10's hello. So, 
10 should advertise 6 as its neighbor. But it does not for about 5 minutes.

Time    Sender  ----Neighbors----
292     6       10      66      7
293     10      66      7
294     6       10      66      7
295     10      66      7
296     6       10      66      7
297     10      66      7
298     6       10      66      7
299     10      66      7
300     6       10      66      7
301     10      66      7
302     6       10      66      7
303     10      66      7
304     6       10      66      7
305     10      66      7
306     6       10      66      7
307     10      66      7
308     6       10      66      7
309     10      66      7
310     6       10      66      7
311     10      66      7
312     6       10      66      7
313     10      66      7
314     6       10      66      7
315     10      66      7
316     6       10      66      7
317     10      66      7
318     6       10      66      7
319     10      66      7

... Pattern continues till time 598.

-- 
Sanjai Narain, Ph.D.
Senior Research Scientist
Information Assurance and Security Department
Telcordia Technologies, Inc. 
1 Telcordia Drive, Room 1N-375
Piscataway, NJ 08854
732 699 2806 (T)
908 337 3636 (M)
narain at research.telcordia.com