LSTOWN-L Archives

LISTSERV List Owners' Forum

LSTOWN-L

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Eric Thomas <[log in to unmask]>
Tue, 14 Jul 1992 15:27:01 +0200
text/plain (40 lines)
Well, I don't think  any policy based on changing the  limit from 300k to
some  other value  is going  to  work, apart  from  the fact  that it  is
something that would require yearly  updates. We have countries connected
at sub-9.6k speed, that is all  they can afford and they should certainly
be allowed to choose which files they  want to let through and which ones
they  want to  trash, given  that there  isn't enough  bandwidth to  send
everything. On the other hand, there is  no reason two US users should be
limited to something as ridiculous as 300k. So even though a common limit
would make  life easier for  servers like  LISTSERV, which could  then be
taught not to send larger files to  remote users, I'm afraid it is simply
not a realistic goal.
 
In EARN there has never been  any limit. Initially this irritated me very
much,  as high  energy  physicists were  happily sending  million-records
files on  a regular basis which  statistically never made it  through, as
the  lines never  stayed up  long enough  to send  that much  data (we're
talking  about several  hours). Since  there was  no limit,  we were  not
allowed  to get  rid  of them  unless  there was  an  emergency, so  what
happened  is  that  we just  had  to  wait  until  we had  an  emergency.
Fortunately, it didn't take very long:  the more the files failed to make
it through,  the more  the physicists got  impatient and  sent additional
copies, to "push" the ones further  up the queue (exact same technique as
for batch jobs).  Eventually the spool reached 90%, we  had an emergency,
and we were able to legally dispose of the trash - until next time. Not a
satisfactory  situation, since  the files  kept using  bandwidth until  a
carrier drop occured and the transfer was aborted.
 
But now that we have a reasonably fast backbone, I am happy that there is
no limit, since it means bureaucratic  sites won't trash my 300k files on
the basis that they  might use up 1 or even 2 seconds  of a poor T1 line.
The limit really has  to be based on the speed of the  lines and the size
of the spool  at the hub nodes,  or it won't satisfy  everyone. You could
perhaps agree on  a limit for the  "high speed" part of  the network, but
you would have  to allow the people operating the  hub nodes that connect
"low-speed" areas (typically countries) to make their own selection based
on local  policy. What you could  probably specify is how  the sender and
recipient should be notified if a file is purged.
 
  Eric

ATOM RSS1 RSS2