Re: [Gzz] attacking gisp peg comments

[Tuomas Lukka]

On Wed, Jun 11, 2003 at 03:08:21PM +0300, Hermanni Hyytiälä wrote:
> 
> On Wed, 2003-06-11 at 13:37, Tuomas Lukka wrote:
> 
> > > We expect that GISP's fault tolerance is at least at the same level as 
> > > Chord_'s 
> > > fault tolerace since GISP's routing table is based on Chord's routing 
> > > table.
> > > 
> > > Chord's general properties:
> > > - O(log^2 n) messages are required to join/leave operations 
> > > - O(log n) lookup efficiency
> > > - Routing table maintains information about O(log n) peers
> > > - Routing table requires information about O(log n) of other peers
> > >   of *efficient* routing, but performance degrades gracefully
> > >   when that information is out of date 
> > > - Only one piece of information per peer need to be corect in 
> > >   order to guarantee correct (though slow) routing queries  
> > 
> > Do you *KNOW* whether this works for GISP with dumb peers or not?
> 
> No I don't. The problem is that we don't know how similar GISP's routing
> table *really* (and overlay network) is to Chord's one.
> 
> That's why I sent an email to Daishi earlier today. 

Umm, I think you send too many emails without first trying to find
out yourself (as in the DHT/DSHT thing). Emailing the author
should not be your *FIRST* option, before you even understand the problem
well.  

Personally, I like it when people ask me things but I don't like
it when unknown people ask me uninformed questions. Both "un"s are needed:
it's ok when people I know ask me uninformed questions - then I'll do my best
to make them informed. Also, when unknown people ask me informed questions
it's a delight to answer. But the unknown-uninformed combination is not nice -
it's something that easily leads to "death of the internet", since LOTS
of unknown people could ask the author of one piece of software uninformed 
questions, which they should have asked from someone near to them.

The source is there and is fairly small - why don't you read and understand
it instead?

Then you could make informed answers instead of just echoing what other people
say even which is especially harmful if they were talking about completely 
different concepts using the same names.

> > > Eventually, we will compare (and validate) our test results with the 
> > > Chord's 
> > > fault tolerance properties.
> > 
> > Not eventually.
> > 
> > YOU NEED TO MAKE A REAL INFORMATIVE STATEMENT ABOUT WHAT THE RESULTS
> > WILL BE THAT COULD BE RIGHT OR WRONG.
> 
> This PEG (and this part of the document) is not ready yet... 
> 
> OTOH, is that obvious that the "RIGHT" results are (almost) similar
> to Chord's results which are listed in the PEG already, e.g., 
> "Result: 20% of lookups fail, when 20% of peers are failed" ?
> 
> Hm...so I should *emphasize* (i.e., repeat) that we expect similar
> results as Chord has gained in the paper, or...?

No, you should not say "expect similar" - it's far too vague.
You should make explicit statements as to what the results will
be.

Stating

        If fraction f of the peers are dumb, then fraction f of the lookups
        will fail at all network sizes.

would be ok. This is an accurate, measurable statement which we can then
start seeing whether it's right or wrong.

Whether it *is* right or wrong, depends of course on the exact behaviour
of GISP which it's your job to find out about (and not emailing the author
except as last recourse).

        Tuomas