LARA

Question 1:

  • Does “Node Ordering Constraint” bug reported in paper works for more than 3 nodes?
  • Answer: Maybe, but this particular scenario does not work for large rings.
  • Description: This figure show the scenario of that bug which is described in paper:

Chord ordering bug

  • However, as described in the bellow figure, it requires that a new successor of $y$ (i.e. $z$) to be in predecessor neighborhood of $x$ which implies a small ring.

Chord ordering bug

Question 2:

  • How long does it take to find ordering bug by exhaustive search in MaceMC, given only 3 node are running?
  • Answer: 7 Hours at depth 16
  • This is the log file: Chord-Q2-LogFile

Question 3:

  • What is the happening rate of Chord bugs?
  • Answer: 37 ordering bug is reported during 62100s for 25 nodes.
  • Notice that modelchecker was runnning per 80 minues, so the actual number of bugs should be like 1000