[cs63201] Some question on AOS

Mikhail Nesterenko mikhail at cs.kent.edu
Wed Oct 9 00:27:30 EDT 2013


> 
> Could you please answer for the following questions?
> 
> Q1.  causality(->) is a irreflexive partial order (i.e. the relation is
> transitive and antisymmetric). how causality(->) anti-symmetric?  I mean is
> it possible to have a -> b and b -> a? if yes what does it mean?

no, irreflexive means that no element can be causally related to
    itself
and anti-symmetric means that if A precedes B then B cannot
    precede A

> 
> Q2.  What is the easy way to count( or find all computations) computations
> in a timing diagram?

Hmm, good question. I would say a permutation of all concurrent events.


> 
> Q3. In Matrix clocks  - does not math R2 with what you have explained with
> example in class (Which one we should follow). Following is the R2 in slide
> and I have added bold text where I found difference.



> R2: attach the whole matrix to each outgoing message, when message received
> from *Pj* update matrix clock as follows
> 
> *mti*[*i,k*] := *max*(*mti*[*i,k*], *mtmsg*[*i,j*]  *this* *should be mtmsg[
> j,k])* for 1≤ *k* ≤ *n* ??? synchronize vector clocks of *Pi* and *Pj *
> 
> * mti*[*i,i*] := *maxk*(*mti*[*i,k*]) ??? synchronize local counter *// you
> didn't consider this line in class *

Right, the above line is incorrect. I removed it from the slide.

> 
> *mti*[*k,l*] := *max*(*mti*[*k,l*], *mtmsg*[*k,l*]) for 1≤ *k,l* ≤ *n* ???
> update the rest of info
> 
> execute R1
> 
> 
> Q4. I need hints about how is real time/delay modeled?

Well, the claim of the asynchronous computation model is that for any
particular real-time based behavior: channel delay, process being slow
compared to the other, etc. there is a computation that models this
behavior. Thus, considering all the possible computations, we can
study program behavior without needing to deal with physical time
(which complicates reasoning about programs)


> Q5. In "tree algorithm", should we always consider for all statement in the
> algorithm you mention in slide?  Or actual tree algorithm is excluding that
> statement?

That depends on the question. Gerard Tel, as well as myself, first
introduce the part of the algorithm where only two processes decide
and then explain the complete algorithm.

Thanks,
--
Mikhail


More information about the cs63201 mailing list