RE: [squid-users] Delay pools explained ??

From: Henrik Nordstrom <hno@dont-contact.us>
Date: 05 Dec 2002 12:17:36 +0100

tor 2002-12-05 klockan 11.37 skrev Rabie van der Merwe:
> What would you then do if say you had two class B's that you wanted to
> rate limit, can you use the same pool, or would you have to create a
> seperate pool per 16bit network? Or say you had 3 24bit networks would
> you then say eiter create one class 3 pool for all 3 or 3 class 2 pools
> one for each network?

If you have more than one class B networks where the last 16 bits may
overlap significantly then you should create multiple class 3 pools. If
the same last 16 bits are very unlikely to be used concurrently in more
than one network then you may get by with only one pool.

For the multiple 24bit networks it is up to you if you create multiple
class 2 pools or one large class 3 pool.

Creating multiple pools is needed if you want to apply different
settings to different networks.

class 2 uses the last 8 bits of the client IP address as "user
identity".

class 3 uses the last 16 bits of the client IP address as "user
identity".

And in addition to the aggregate and per user limits class 3 pools also
have network pool limits which works like the aggregate pool limits but
apply to the 24 bit networks. A class 3 pool has 256 networks of 256
client IP addresses, based on the last 16 bits of the client IP address.

Regards
Henrik
Received on Thu Dec 05 2002 - 04:17:47 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:11:51 MST