summaryrefslogtreecommitdiffstats
path: root/man/man8/tc-fq_pie.8
blob: 457a56bb12a5ae83efc4fe9be3d8b74fec02cfb6 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
.TH FQ-PIE 8 "23 January 2020" "iproute2" "Linux"

.SH NAME

FQ-PIE - Flow Queue Proportional Integral controller Enhanced

.SH SYNOPSIS

.B tc qdisc ... fq_pie
[ \fBlimit\fR PACKETS ] [ \fBflows\fR NUMBER ]
.br
                    \
[ \fBtarget\fR TIME ] [ \fBtupdate\fR TIME ]
.br
                    \
[ \fBalpha\fR NUMBER ] [ \fBbeta\fR NUMBER ]
.br
                    \
[ \fBquantum\fR BYTES ] [ \fBmemory_limit\fR BYTES ]
.br
                    \
[ \fBecn_prob\fR PERENTAGE ] [ [\fBno\fR]\fBecn\fR ]
.br
                    \
[ [\fBno\fR]\fBbytemode\fR ] [ [\fBno_\fR]\fBdq_rate_estimator\fR ]

.SH DESCRIPTION
FQ-PIE (Flow Queuing with Proportional Integral controller Enhanced) is a
queuing discipline that combines Flow Queuing with the PIE AQM scheme. FQ-PIE
uses a Jenkins hash function to classify incoming packets into different flows
and is used to provide a fair share of the bandwidth to all the flows using the
qdisc. Each such flow is managed by the PIE algorithm.

.SH ALGORITHM
The FQ-PIE algorithm consists of two logical parts: the scheduler which selects
which queue to dequeue a packet from, and the PIE AQM which works on each of the
queues. The major work of FQ-PIE is mostly in the scheduling part. The
interaction between the scheduler and the PIE algorithm is straight forward.

During the enqueue stage, a hashing-based scheme is used, where flows are hashed
into a number of buckets with each bucket having its own queue. The number of
buckets is configurable, and presently defaults to 1024 in the implementation.
The flow hashing is performed on the 5-tuple of source and destination IP
addresses, port numbers and IP protocol number. Once the packet has been
successfully classified into a queue, it is handed over to the PIE algorithm
for enqueuing. It is then added to the tail of the selected queue, and the
queue's byte count is updated by the packet size. If the queue is not currently
active (i.e., if it is not in either the list of new or the list of old queues)
, it is added to the end of the list of new queues, and its number of credits
is initiated to the configured quantum. Otherwise, the queue is left in its
current queue list.

During the dequeue stage, the scheduler first looks at the list of new queues;
for the queue at the head of that list, if that queue has a negative number of
credits (i.e., it has already dequeued at least a quantum of bytes), it is given
an additional quantum of credits, the queue is put onto the end of the list of
old queues, and the routine selects the next queue and starts again. Otherwise,
that queue is selected for dequeue again. If the list of new queues is empty,
the scheduler proceeds down the list of old queues in the same fashion
(checking the credits, and either selecting the queue for dequeuing, or adding
credits and putting the queue back at the end of the list). After having
selected a queue from which to dequeue a packet, the PIE algorithm is invoked
on that queue.

Finally, if the PIE algorithm does not return a packet, then the queue must be
empty and the scheduler does one of two things:

If the queue selected for dequeue came from the list of new queues, it is moved
to the end of the list of old queues. If instead it came from the list of old
queues, that queue is removed from the list, to be added back (as a new queue)
the next time a packet arrives that hashes to that queue. Then (since no packet
was available for dequeue), the whole dequeue process is restarted from the
beginning.

If, instead, the scheduler did get a packet back from the PIE algorithm, it
subtracts the size of the packet from the byte credits for the selected queue
and returns the packet as the result of the dequeue operation.

.SH PARAMETERS
.SS limit
It is the limit on the queue size in packets. Incoming packets are dropped when
the limit is reached. The default value is 10240 packets.

.SS flows
It is the number of flows into which the incoming packets are classified. Due
to the stochastic nature of hashing, multiple flows may end up being hashed
into the same slot. Newer flows have priority over older ones. This
parameter can be set only at load time since memory has to be allocated for
the hash table. The default value is 1024.

.SS target
It is the queue delay which the PIE algorithm tries to maintain. The default
target delay is 15ms.

.SS tupdate
It is the time interval at which the system drop probability is calculated.
The default is 15ms.

.SS alpha
.SS beta
alpha and beta are parameters chosen to control the drop probability. These
should be in the range between 0 and 32.

.SS quantum
quantum signifies the number of bytes that may be dequeued from a queue before
switching to the next queue in the deficit round robin scheme.

.SS memory_limit
It is the maximum total memory allowed for packets of all flows. The default is
32Mb.

.SS ecn_prob
It is the drop probability threshold below which packets will be ECN marked
instead of getting dropped. The default is 10%. Setting this parameter requires
\fBecn\fR to be enabled.

.SS \fR[\fBno\fR]\fBecn\fR
It has the same semantics as \fBpie\fR and can be used to mark packets
instead of dropping them. If \fBecn\fR has been enabled, \fBnoecn\fR can
be used to turn it off and vice-a-versa.

.SS \fR[\fBno\fR]\fBbytemode\fR
It is used to scale drop probability proportional to packet size
\fBbytemode\fR to turn on bytemode, \fBnobytemode\fR to turn off
bytemode. By default, \fBbytemode\fR is turned off.

.SS \fR[\fBno_\fR]\fBdq_rate_estimator\fR
\fBdq_rate_estimator\fR can be used to calculate queue delay using Little's
Law, \fBno_dq_rate_estimator\fR can be used to calculate queue delay
using timestamp. By default, \fBdq_rate_estimator\fR is turned off.

.SH EXAMPLES
# tc qdisc add dev eth0 root fq_pie
.br
# tc -s qdisc show dev eth0
.br
qdisc fq_pie 8001: root refcnt 2 limit 10240p flows 1024 target 15.0ms tupdate
16.0ms alpha 2 beta 20 quantum 1514b memory_limit 32Mb ecn_prob 10
 Sent 159173586 bytes 105261 pkt (dropped 24, overlimits 0 requeues 0)
 backlog 75700b 50p requeues 0
  pkts_in 105311 overlimit 0 overmemory 0 dropped 24 ecn_mark 0
  new_flow_count 7332 new_flows_len 0 old_flows_len 4 memory_used 108800

# tc qdisc add dev eth0 root fq_pie dq_rate_estimator
.br
# tc -s qdisc show dev eth0
.br
qdisc fq_pie 8001: root refcnt 2 limit 10240p flows 1024 target 15.0ms tupdate
16.0ms alpha 2 beta 20 quantum 1514b memory_limit 32Mb ecn_prob 10
dq_rate_estimator
 Sent 8263620 bytes 5550 pkt (dropped 4, overlimits 0 requeues 0)
 backlog 805448b 532p requeues 0
  pkts_in 6082 overlimit 0 overmemory 0 dropped 4 ecn_mark 0
  new_flow_count 94 new_flows_len 0 old_flows_len 8 memory_used 1157632

.SH SEE ALSO
.BR tc (8),
.BR tc-pie (8),
.BR tc-fq_codel (8)

.SH SOURCES
RFC 8033: https://tools.ietf.org/html/rfc8033

.SH AUTHORS
FQ-PIE was implemented by Mohit P. Tahiliani. Please report corrections to the
Linux Networking mailing list <netdev@vger.kernel.org>.