Re: WG ACTION: 2 weeks to discuss [LL61] Remove initial wait

I don't think the new text correctly captures either of the behaviors we've 
discussed - I don't see where an initial delay is specified.

I think this text is correct:

    When ready to begin probing, the host should then wait for a random
    time interval selected uniformly in the range 0 to INITIAL_DELAY
    seconds, and should then send NUM_PROBES probe packets, spaced
    randomly, PROBE_MIN to PROBE_MAX seconds apart.

where INITIAL_DELAY is defined somewhere to be 1.

- Ralph

At 01:07 AM 5/5/2004 +0200, Erik Guttman wrote:
>Please post discussion of this issue to the mailing list over the next two 
>weeks
>ending May 18, 2004.  In order to accept this issued, we will need a strong WG
>consensus given that this is very late in the process.
>
>Please see http://www.drizzle.org/~aboba/ZEROCONF/issues.html for a list of
>current issues and their status.
>
>[LL61]
>
>Description of Issue:           Remove initial wait
>Submitter Name:                 Stuart Cheshire
>Submitter Email Address:        [email protected]
>Date first submitted:           04 May 04
>Reference:                      LL12
>Comment Type ['t'ech|'e'dit]:   t
>Prio ['S' Must|1 should|2 may]: S
>Section:                        2.2.1
>Rationale/Explanation:
>Lengthy Description:
>
>[Stuart]
>
>>    When ready to begin probing, the host should then wait for a random
>>    time interval selected uniformly in the range PROBE_MIN 0
>>    seconds, and should then send NUM_PROBES probe packets, spaced
>>    randomly, PROBE_MIN to PROBE_MAX seconds apart.
>
>Why "wait for a random time interval selected uniformly in the range
>PROBE_MIN to PROBE_MAX"? What's the benefit of enforcing an initial
>one-second delay? It just slows things down.
>
>[Erik]
>
>This delay was intended to stop a set of hosts from beginning at the
>same time in a 'LAN power up' situation.  This text has passed all
>reviews and numerous WG issues to revise and hone it.
>
>[Stuart]
>
>I was not asking about the [0,1] random interval. That's been there since
>draft-05.
>
>I was asking about why it is now 1 + [0,1]. What's the extra fixed
>one-second delay for? What is achieved by making the process uniformly
>take a second longer than it should?
>
>[Erik]
>
>     Hmm.  Reviewing all records, I can't see how this entered the doc.
>     I don't have time for archeology to find out when it entered.  I
>     don't see why waiting an extra second helps, except to wait for
>     network infrastructure to come up (see ll12).
>
>Requested Change:
>
>Text was:
>
>    When ready to begin probing, the host should then wait for a random
>    time interval selected uniformly in the range PROBE_MIN to PROBE_MAX
>    seconds, and should then send NUM_PROBES probe packets, spaced
>    randomly, PROBE_MIN to PROBE_MAX seconds apart.
>
>Text becomes:
>
>    When ready to begin probing, the host should send NUM_PROBES probe
>    packets, spaced randomly, PROBE_MIN to PROBE_MAX seconds apart.