QoS

Unanswered Question
Dec 18th, 2008

guy i am very new to QoS just to make my self understand what is the difference between ip precedence and DSCP??? i guess both are used for QoS what is the difference secondly when we make a class-map for example class-map cc.....does it mean that match all....as i have seen another class-map match-all cc.....what is the difference between them....thanks alot for looking one more thing i have created class-map cc on a router and when i do show class-map it shows match-all staement in it......in running config just wanted to know why we write it and when we dont write it....thanks bud

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Joseph W. Doherty Thu, 12/18/2008 - 07:31

The difference between IP Predence and DSCP is how the IP header's ToS byte bits are allocated and their purpose. Basically IP Predence is based on an older RFC than the DSCP RFC.

Either can be used, or not, for QoS purposes. There are suggestions for how either might be used, but no, I believe, requirements.

How DSCP might be used was very open ended and there have been follow-on RFCs addressing possible usage. The more important ones concern assured forwarding (AFxy) and expedited forwarding (EF), which you'll likely bump into.

QoS is one of those subjects, that I don't think is really difficult to understand, but requires knowing lots of information for it to be well understood.

You might try first reading some Wiki articles on the subject along with some Cisco infomation other than their IOS reference manuals. A good book on the subject can be a great aid.

Craig Norborg Thu, 12/18/2008 - 07:51

You can use either in the same way if you want, although DSCP is more current and standard. IP Precedence is somewhat a predecessor to DSCP and DSCP is for the most part backwards compatible.

From a technical viewpoint, IP Precedence is the top 3 (bits 5 - 7) of the TOS Byte. DSCP is the top 6 (bits 2 - 7) of the TOS Byte.

johnlloyd_13 Fri, 12/19/2008 - 00:21

answering your second question regarding class-map config. the difference between match-all and match-any is their logical operation. in other words, match-all (which is the default) operates as a logical AND. "all" of the match statements (under config-cmap subconfig) must be met for packets to be grouped under this class. on the other hand, match-any is like a logical OR. "at least one" match statements must be met for packets to be grouped under this class. we write class maps first (following MQC) to group/identify/classify traffic we wanted to apply QoS.

Actions

This Discussion