Re: Delay pools

[Note: message from David Luyer, using an alternate e-mail address, since
I get mail bounces which look like spam blocks from rlim@playnet.com.au
whenever I e-mail the list from my normal e-mail address of
luyer@ucs.uwa.edu.au, and this thread is a response to his post so using
that address would be pretty pointless]

David J Wolley wrote:

> It delays the rate at which the user can receive data, [...]

This is _wrong_. In one message I accidentally made it sound a bit like
this.

Delay pools do _not_ affect already cached objects. They do _not_ affect
queries sent to a peer marked with the no_delay tag (eg, local peer hits).
They also are cumulative when multiple requestors hop on an object (ie, if
two people have a 384 byte/sec 'individual' limit and both of them
simultaneously request the same object, the permitted rate will be
768 byte/sec).

It is correct that it limits data rates on a per-IP basis, ie, per-user,
if appropriately configured, however to say it limits the rate at which a
user can receive data is wrong because it implies it limits the transfer
rate of already cached objects, something it definitely does not do.

'rlim' - if you get this message, read the mailing list archive, I answered
the rest of your questions already and have also explained delay pools many
times previously which your mailer bounced.

David.
Received on Thu Feb 25 1999 - 18:56:46 MST