RE: Sort of OT: Hardware Queues on Serial Interfaces

From: Scott Morris (swm@emanon.com)
Date: Tue Jun 08 2004 - 23:18:42 GMT-3


No. There is a single priority queue (LLQ). You can map multiple classes
into this single queue, thus expanding its size, but that particular single
queue is handled in a FIFO manner. Just before everything else.

HTH,

 
Scott Morris, CCIE4 (R&S/ISP-Dial/Security/Service Provider) #4713, CISSP,
JNCIP, et al.
IPExpert CCIE Program Manager
IPExpert Sr. Technical Instructor
swm@emanon.com/smorris@ipexpert.net
http://www.ipexpert.net
 
 

-----Original Message-----
From: GAD [mailto:gad@gad.net]
Sent: Tuesday, June 08, 2004 10:16 PM
To: Scott Morris
Cc: ccielab@groupstudy.com
Subject: RE: Sort of OT: Hardware Queues on Serial Interfaces

Hmm... then I misunderstood.

Let me refine the question then:

Is it possible to prioritize different types of traffic *WITHIN* the
priority queue?

Say for example you've got calls that make money and calls that don't.
Both need strict priority, but can you prioritize one over the other?

For example:

class-map match-all MoneyMaker
  match input-interface FastEthernet0/1
!
class-map match-all Admin
  match input-interface FastEthernet0/0
!
!
policy-map Calls
  class MoneyMaker
    priority 2000000
  class Admin
    priority 500000

Will this service the queue more "often" for the MoneyMaker class than the
Admin class, and more importantly, will it negatively affect either?

It doesn't look or sound like a good idea to me, by I in part need to
understand it, and in part need to counter someone's argument.

Thanks,

GAD

On Tue, 8 Jun 2004, Scott Morris wrote:

> The LLQ gets serviced in strict priority LIKE the broadcast queue on a
> serial interface, but everything from LLQ flows into the "normal"
> queue just like other queues do. The hardward broadcast queue is
> reserved for broadcasts, pseudo-broadcasts and L2 multicast stuff.
>
> HTH,
>
>
> Scott Morris, CCIE4 (R&S/ISP-Dial/Security/Service Provider) #4713,
> CISSP,



This archive was generated by hypermail 2.1.4 : Sat Jul 03 2004 - 19:40:36 GMT-3