Re: QoS default-class - oversubscription - VERIFICATION in LAB

From: Pierre-Alex (paguanel@hotmail.com)
Date: Sun Jun 04 2006 - 10:17:59 ART


Hi Group,

Here are some lab results on oversubscription and the default-class verifying
Cisco Theory.

It was very useful to me in clarifying things.

1- RIP, OSPF and EIGRP update traffic will not be dropped even if 100% of
link is reserved to other classes and these classes use their allocations.

2- When saturating a link, there is 1K of traffic that cannot be used -
reseverved by the IOS .

3- BGP will not be able to function because of not having pak_priority set if
the link is saturated.

For details explainations you can read the summary below (from Cisco Doc).

"The RIP and OSPF (and EIGRP) routing processes running on a router's core CPU
mark all traffic they originate with both IP precedence 6 and pak_priority. In
contrast, the Border Gateway Protocol (BGP) instructs TCP to mark its traffic
with IP precedence 6, but does not set pak_priority.

RSP-based QoS and other platforms, including the Cisco 7200, 3600, 2600 series
places pak_priority traffic in a separate set of queues other than
class-default and marks such messages with a special weight value (currently
1024)."

http://www.cisco.com/en/US/tech/tk543/tk544/technologies_tech_note09186a00800
94612.shtml

Kind regards,

Pierre



This archive was generated by hypermail 2.1.4 : Sat Jul 01 2006 - 07:57:31 ART