RE: Does Precedence 5 always equal DCSP CS5?

From: Scott Morris (swm@emanon.com)
Date: Sun Nov 16 2003 - 15:13:30 GMT-3


You have to look back at the details within RFC791 to get a true feeling
for why DSCP 46 is more appropriate for voice traffic. Those other
three bits, while not part of IP Precedence, still had value...

IP Prec 5 by itself translates to DSCP 40, plain and simple.

However:

 Type of Service: 8 bits

    The Type of Service provides an indication of the abstract
    parameters of the quality of service desired. These parameters are
    to be used to guide the selection of the actual service parameters
    when transmitting a datagram through a particular network. Several
    networks offer service precedence, which somehow treats high
    precedence traffic as more important than other traffic (generally
    by accepting only traffic above a certain precedence at time of high
    load). The major choice is a three way tradeoff between low-delay,
    high-reliability, and high-throughput.

      Bits 0-2: Precedence.
      Bit 3: 0 = Normal Delay, 1 = Low Delay.
      Bits 4: 0 = Normal Throughput, 1 = High Throughput.
      Bits 5: 0 = Normal Relibility, 1 = High Relibility.
      Bit 6-7: Reserved for Future Use.

So Voice traffic should not only be classified as IP Prec 5, but also
Low Delay (bit 4 = 1) and High Throughput (bit 5 = 1) giving us the
total package of 101110 or 46 in DSCP terms.

So it really isn't so much that anything is changing other than the
perspective of what bits we are measuring where. The bits should really
still be the same regardless of whether our router is DSCP-capable or
not, but we then have more granularity to play with in order to make
this wonderful pain-in-the-rear technology of voice play nicely on the
networks!

Cheers,

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

-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Brian McGahan
Sent: Saturday, November 15, 2003 7:52 PM
To: 'Michael Snyder'; ccielab@groupstudy.com
Subject: RE: Does Precedence 5 always equal DCSP CS5?

Michael,

        As a side note, setting IP precedence on a local POTS peer is no
longer supported. However you can still set it using the appropriate
DSCP -> IP Prec mappings. Since DSCP implicitly overlaps IP precedence,
DSCP is the preferred way to color traffic.

Router(config-dial-peer)#ip prec?
% Unrecognized command
Router(config-dial-peer)#ip qos dscp ?
  <0-63> DSCP value
  af11 Sets DSCP to assured forwarding (af11) bit pattern 001010
  af12 Sets DSCP to assured forwarding (af12) bit pattern 001100
  af13 Sets DSCP to assured forwarding (af13) bit pattern 001110
  af21 Sets DSCP to assured forwarding (af21) bit pattern 010010
  af22 Sets DSCP to assured forwarding (af22) bit pattern 010100
  af23 Sets DSCP to assured forwarding (af23) bit pattern 010110
  af31 Sets DSCP to assured forwarding (af31) bit pattern 011010
  af32 Sets DSCP to assured forwarding (af32) bit pattern 011100
  af33 Sets DSCP to assured forwarding (af33) bit pattern 011110
  af41 Sets DSCP to assured forwarding (af41) bit pattern 100010
  af42 Sets DSCP to assured forwarding (af42) bit pattern 100100
  af43 Sets DSCP to assured forwarding (af43) bit pattern 100110
  cs1 Sets DSCP to class selector codepoint 1 (precedence 1)
  cs2 Sets DSCP to class selector codepoint 2 (precedence 2)
  cs3 Sets DSCP to class selector codepoint 3 (precedence 3)
  cs4 Sets DSCP to class selector codepoint 4 (precedence 4)
  cs5 Sets DSCP to class selector codepoint 5 (precedence 5)
  cs6 Sets DSCP to class selector codepoint 6 (precedence 6)
  cs7 Sets DSCP to class selector codepoint 7 (precedence 7)
  default Sets DSCP to default bit pattern of 000000
  ef Sets DSCP to expedited forwarding bit pattern 101110

HTH,

Brian McGahan, CCIE #8593
bmcgahan@internetworkexpert.com

Internetwork Expert, Inc.
http://www.InternetworkExpert.com
Toll Free: 877-224-8987
Direct: 708-362-1418 (Outside the US and Canada)

> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
Of
> Michael Snyder
> Sent: Saturday, November 15, 2003 9:14 AM
> To: ccielab@groupstudy.com
> Cc: 'Donny MATEO'
> Subject: RE: Does Precedence 5 always equal DCSP CS5?
>
> Hold on,
>
> DCSP CS5 doesn't equal DCSP EF
>
> Assuming the TOS bits are zero, Precedence 5 should equal DCSP 40.
>
> Using TOS, maybe you could get to EF(46).
>
> I've never played with the TOS bits.
>
> -----Original Message-----
> From: Donny MATEO [mailto:donny.mateo@sg.ca-indosuez.com]
> Sent: Saturday, November 15, 2003 10:34 AM
> To: Michael Snyder
> Cc: ccielab@groupstudy.com; nobody@groupstudy.com
> Subject: Re: Does Precedence 5 always equal DCSP CS5?
>
> I seem to remember that DSCP is has another 3 bits addition to the ip
> precedence bits. Those 3 bits are the factor determining the drop
preference....(actually
>
> only 2 is used if I recall correctly). So yeah precedence 5 (EF), but
> what kind of drop characteristic ?...... I'm not sure if the
> implementation on
> IOS QOS coding actually take heed of those extra bits in DSCP, but who
> knows.
>
> Donny
>
>



This archive was generated by hypermail 2.1.4 : Fri Dec 12 2003 - 12:29:12 GMT-3