| 1 |
.TH FQ_CoDel 8 "4 June 2012" "iproute2" "Linux" |
| 2 |
.SH NAME |
| 3 |
CoDel \- Fair Queuing (FQ) with Controlled Delay (CoDel) |
| 4 |
.SH SYNOPSIS |
| 5 |
.B tc qdisc ... fq_codel |
| 6 |
[ |
| 7 |
.B limit |
| 8 |
PACKETS ] [ |
| 9 |
.B flows |
| 10 |
NUMBER ] [ |
| 11 |
.B target |
| 12 |
TIME ] [ |
| 13 |
.B interval |
| 14 |
TIME ] [ |
| 15 |
.B quantum |
| 16 |
BYTES ] [ |
| 17 |
.B ecn |
| 18 |
| |
| 19 |
.B noecn |
| 20 |
] [ |
| 21 |
.B ce_threshold |
| 22 |
TIME ] [ |
| 23 |
.B ce_threshold_selector |
| 24 |
VALUE/MASK ] [ |
| 25 |
.B memory_limit |
| 26 |
BYTES ] |
| 27 |
|
| 28 |
.SH DESCRIPTION |
| 29 |
FQ_Codel (Fair Queuing Controlled Delay) is queuing discipline that combines Fair |
| 30 |
Queuing with the CoDel AQM scheme. FQ_Codel uses a stochastic model to classify |
| 31 |
incoming packets into different flows and is used to provide a fair share of the |
| 32 |
bandwidth to all the flows using the queue. Each such flow is managed by the |
| 33 |
CoDel queuing discipline. Reordering within a flow is avoided since Codel |
| 34 |
internally uses a FIFO queue. |
| 35 |
|
| 36 |
.SH PARAMETERS |
| 37 |
.SS limit |
| 38 |
has the same semantics as |
| 39 |
.B codel |
| 40 |
and is the hard limit on the real queue size. |
| 41 |
When this limit is reached, incoming packets are dropped. Default is 10240 |
| 42 |
packets. |
| 43 |
|
| 44 |
.SS memory_limit |
| 45 |
sets a limit on the total number of bytes that can be queued in this FQ-CoDel |
| 46 |
instance. The lower of the packet limit of the |
| 47 |
.B limit |
| 48 |
parameter and the memory limit will be enforced. Default is 32 MB. |
| 49 |
|
| 50 |
|
| 51 |
.SS flows |
| 52 |
is the number of flows into which the incoming packets are classified. Due to |
| 53 |
the stochastic nature of hashing, multiple flows may end up being hashed into |
| 54 |
the same slot. Newer flows have priority over older ones. This parameter can be |
| 55 |
set only at load time since memory has to be allocated for the hash table. |
| 56 |
Default value is 1024. |
| 57 |
|
| 58 |
.SS target |
| 59 |
has the same semantics as |
| 60 |
.B codel |
| 61 |
and is the acceptable minimum |
| 62 |
standing/persistent queue delay. This minimum delay is identified by tracking |
| 63 |
the local minimum queue delay that packets experience. Default value is 5ms. |
| 64 |
|
| 65 |
.SS interval |
| 66 |
has the same semantics as |
| 67 |
.B codel |
| 68 |
and is used to ensure that the measured minimum delay does not become too stale. |
| 69 |
The minimum delay must be experienced in the last epoch of length |
| 70 |
.BR interval . |
| 71 |
It should be set on the order of the worst-case RTT through the bottleneck to |
| 72 |
give endpoints sufficient time to react. Default value is 100ms. |
| 73 |
|
| 74 |
.SS quantum |
| 75 |
is the number of bytes used as 'deficit' in the fair queuing algorithm. Default |
| 76 |
is set to 1514 bytes which corresponds to the Ethernet MTU plus the hardware |
| 77 |
header length of 14 bytes. |
| 78 |
|
| 79 |
.SS ecn | noecn |
| 80 |
has the same semantics as |
| 81 |
.B codel |
| 82 |
and can be used to mark packets instead of dropping them. If |
| 83 |
.B ecn |
| 84 |
has been enabled, |
| 85 |
.B noecn |
| 86 |
can be used to turn it off and vice-a-versa. Unlike |
| 87 |
.B codel, ecn |
| 88 |
is turned on by default. |
| 89 |
|
| 90 |
.SS ce_threshold |
| 91 |
sets a threshold above which all packets are marked with ECN Congestion |
| 92 |
Experienced. This is useful for DCTCP-style congestion control algorithms that |
| 93 |
require marking at very shallow queueing thresholds. |
| 94 |
|
| 95 |
.SS ce_threshold_selector |
| 96 |
sets a filter so that the |
| 97 |
.B ce_threshold |
| 98 |
feature is applied to only a subset of the traffic seen by the qdisc. If set, the MASK value |
| 99 |
will be applied as a bitwise AND to the diffserv/ECN byte of the IP header, and only if the |
| 100 |
result of this masking equals VALUE, will the |
| 101 |
.B ce_threshold |
| 102 |
logic be applied to the packet. |
| 103 |
|
| 104 |
.SS drop_batch |
| 105 |
sets the maximum number of packets to drop when |
| 106 |
.B limit |
| 107 |
or |
| 108 |
.B memory_limit |
| 109 |
is exceeded. Default value is 64. |
| 110 |
|
| 111 |
.SH EXAMPLES |
| 112 |
#tc qdisc add dev eth0 root fq_codel |
| 113 |
.br |
| 114 |
#tc -s qdisc show |
| 115 |
.br |
| 116 |
qdisc fq_codel 8002: dev eth0 root refcnt 2 limit 10240p flows 1024 quantum 1514 |
| 117 |
target 5.0ms interval 100.0ms ecn |
| 118 |
Sent 428514 bytes 2269 pkt (dropped 0, overlimits 0 requeues 0) |
| 119 |
backlog 0b 0p requeues 0 |
| 120 |
maxpacket 256 drop_overlimit 0 new_flow_count 0 ecn_mark 0 |
| 121 |
new_flows_len 0 old_flows_len 0 |
| 122 |
|
| 123 |
#tc qdisc add dev eth0 root fq_codel limit 2000 target 3ms interval 40ms noecn |
| 124 |
.br |
| 125 |
#tc -s qdisc show |
| 126 |
.br |
| 127 |
qdisc fq_codel 8003: dev eth0 root refcnt 2 limit 2000p flows 1024 quantum 1514 |
| 128 |
target 3.0ms interval 40.0ms |
| 129 |
Sent 2588985006 bytes 1783629 pkt (dropped 0, overlimits 0 requeues 34869) |
| 130 |
backlog 0b 0p requeues 34869 |
| 131 |
maxpacket 65226 drop_overlimit 0 new_flow_count 73 ecn_mark 0 |
| 132 |
new_flows_len 1 old_flows_len 3 |
| 133 |
|
| 134 |
|
| 135 |
.SH SEE ALSO |
| 136 |
.BR tc (8), |
| 137 |
.BR tc-codel (8), |
| 138 |
.BR tc-red (8) |
| 139 |
|
| 140 |
.SH AUTHORS |
| 141 |
FQ_CoDel was implemented by Eric Dumazet. This manpage was written |
| 142 |
by Vijay Subramanian. Please report corrections to the Linux Networking |
| 143 |
mailing list <[email protected]>. |