Editing Differentiated services (section)

===Assured Forwarding===

The IETF defines the Assured Forwarding (AF) behavior in {{IETF RFC|2597}} and {{IETF RFC|3260}}.  Assured forwarding allows the operator to provide assurance of delivery as long as the traffic does not exceed some subscribed rate.  Traffic that exceeds the subscription rate faces a higher probability of being dropped if congestion occurs.

The AF behavior group defines four separate AF classes with all traffic within one class having the same priority.  Within each class, packets are given a drop precedence (high, medium or low, where higher precedence means ''more'' dropping). The combination of classes and drop precedence yields twelve separate DSCP encodings from AF11 through AF43 (see table).

{| class="wikitable"
|+ Assured Forwarding behavior group
!Drop<br>probability
! Class 1 !! Class 2 !! Class 3 !! Class 4
|-
! Low
| AF11 (DSCP 10) 001010 || AF21 (DSCP 18) 010010 || AF31 (DSCP 26) 011010 || AF41 (DSCP 34) 100010
|-
! Medium
| AF12 (DSCP 12) 001100 || AF22 (DSCP 20) 010100  || AF32 (DSCP 28) 011100 || AF42 (DSCP 36) 100100
|-
! High
| AF13 (DSCP 14) 001110 || AF23 (DSCP 22) 010110 || AF33 (DSCP 30) 011110 || AF43 (DSCP 38) 100110
|}

Some measure of priority and proportional fairness is defined between traffic in different classes. Should congestion occur ''between'' classes, the traffic in the higher class is given priority. Rather than using strict priority queuing, more balanced queue servicing algorithms such as [[fair queuing]] or [[weighted fair queuing]] are likely to be used. If congestion occurs ''within'' a class, the packets with the higher drop precedence are discarded first. Re-marking a packet is sometimes used to increase its drop precedence if a stream's bandwidth exceeds a certain threshold. For example, a stream whose rate is above the Committed Information Rate (CIR) as defined in {{IETF RFC|2697}} causes the stream to be marked with a higher AF drop precedence. This allows the decision as to when to shape the stream to devices further downstream if they encounter congestion. To prevent issues associated with [[tail drop]], more sophisticated drop selection algorithms such as [[random early detection]] are often used.