[chord] Question on Chord: A Scalable Peer-to-peer.....

[Ion Stoica]

Of course, you can also use "to", but the code is just cleaner
using "downto".  If you use "to" instead of "downto" the code
becomes something like:

for i = 1 to m-1
  if ((finger[i].node in (n, id)) && (finger[i+1].node notin (n, id)))
    return finger[i].node;

which is a bit more complex.

Hope this helps,

Ion


진명희 wrote:

> Thank you for your answer~
>
> May I ask one more question?
>
> The question is on Figure 4(The pseudocode...) of 5 page of paper Chord.
> In third paragraph on Figure 4, in the second line of 
> n.closest_preceding_finger(id) procedure "for i=m downto 1" ,
> My question is that if I modify as "for i=1 to m", what's different?
> I think that "for i=m downto 1" is to lookup from faraway to close 
> and "for i=1 to m" is to lookup from close to faraway. Isn't it?
>
> I would like to know why you use "for i=m downto 1" instead of "for 
> i=1 to m".
>
>
> Please, Let me know..
>
> Thanks~
>
> -----원본 메시지-----
> *보낸 사람:* "Ion Stoica" <istoica at cs.berkeley.edu>
> *보낸 날짜:* 2004-01-30 오전 4:59:03
> *받는 사람:* "진명희" <ppuny79 at korea.com>
> *제목:* Re: [chord] Question on Chord: A Scalable Peer-to-peer.....
>
> The keys in Figure 3(b) are arbitrarily chosen. In practice,
> the key (id) of a data item can be computed as a hash (e.g., SHA-1)
> on the data itself, or it can be any unique key assigned by
> applications/users.
>
> Ion
>
>
> 진명희 wrote:
>
> >
> > Thank you for your answer~
> >
> > I have one more question.
> > The question is in Figure 3 (b) on page 5 of the paper Chord.
> > I understood every thing about routing table but the key values..
> > I don't know how the key values are assigned.
> > Please, let me understand it..
> >
> > Thanks, have a nice day~~~
> >
> > -----원본 메시지-----
> > *보낸 사람:* "Frank Dabek" <fdabek at MIT.EDU>
> > *보낸 날짜:* 2004-01-29 오전 6:52:13
> > *받는 사람:* "진명 희" <ppuny79 at korea.com>
> > *제목:* Re: [chord] Question on Chord: A Scalable Peer-to-peer.....
> >
> >
> >The recente efforts to crack the RC5 cipher using a distributed network
> >of volunteers are an example this type of application. Instead of using
> >a central server to distribute parts of the keyspace one could imagine
> >using chord to maintain, in a dencentralized fashion, the range of keys
> >each node should search: i.e. try all of the keys between yourself and
> >your successor.
> >
> >--Frank
> >
> >On Wed, 2004-01-28 at 04:55, 진명희 wrote:
> >> Hi~
> >> May I ask a question on Chord?
> >> I couldn't understand 'Large-Scale Combinatorial Search' on page 3 of
> >> the paper Chord.
> >> Would you give me an example about it?
> >>
> >> Thanks~~~
> >>
> >>
> >>
> >>
> >> ______________________________________________________________________
> >> _______________________________________________
> >> chord mailing list
> >> chord at amsterdam.lcs.mit.edu
> >> https://amsterdam.lcs.mit.edu/mailman/listinfo/chord
> >
> >
> >
> >------------------------------------------------------------------------
> >
> >_______________________________________________
> >chord mailing list
> >chord at amsterdam.lcs.mit.edu
> >https://amsterdam.lcs.mit.edu/mailman/listinfo/chord
> > 
> >
>
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>chord mailing list
>chord at amsterdam.lcs.mit.edu
>https://amsterdam.lcs.mit.edu/mailman/listinfo/chord
>  
>