| 1 |
.TH SKBPRIO 8 "13 August 2018" "iproute2" "Linux" |
| 2 |
.SH NAME |
| 3 |
skbprio \- SKB Priority Queue |
| 4 |
|
| 5 |
.SH SYNOPSIS |
| 6 |
.B tc qdisc ... add skbprio |
| 7 |
.B [ limit |
| 8 |
packets |
| 9 |
.B ] |
| 10 |
|
| 11 |
.SH DESCRIPTION |
| 12 |
SKB Priority Queue is a queueing discipline intended to prioritize |
| 13 |
the most important packets during a denial-of-service ( |
| 14 |
.B DoS |
| 15 |
) attack. The priority of a packet is given by |
| 16 |
.B skb->priority |
| 17 |
, where a higher value places the packet closer to the exit of the queue. When |
| 18 |
the queue is full, the lowest priority packet in the queue is dropped to make |
| 19 |
room for the packet to be added if it has higher priority. If the packet to be |
| 20 |
added has lower priority than all packets in the queue, it is dropped. |
| 21 |
|
| 22 |
Without SKB priority queue, queue length limits must be imposed |
| 23 |
on individual sub-queues, and there is no straightforward way to enforce |
| 24 |
a global queue length limit across all priorities. SKBprio queue enforces |
| 25 |
a global queue length limit while not restricting the lengths of |
| 26 |
individual sub-queues. |
| 27 |
|
| 28 |
While SKB Priority Queue is agnostic to how |
| 29 |
.B skb->priority |
| 30 |
is assigned. A typical use case is to copy |
| 31 |
the 6-bit DS field of IPv4 and IPv6 packets using |
| 32 |
.BR tc-skbedit (8). |
| 33 |
If |
| 34 |
.B skb->priority |
| 35 |
is greater or equal to 64, the priority is assumed to be 63. |
| 36 |
Priorities less than 64 are taken at face value. |
| 37 |
|
| 38 |
SKB Priority Queue enables routers to locally decide which |
| 39 |
packets to drop under a DoS attack. |
| 40 |
Priorities should be assigned to packets such that the higher the priority, |
| 41 |
the more expected behavior a source shows. |
| 42 |
So sources have an incentive to play by the rules. |
| 43 |
|
| 44 |
.SH ALGORITHM |
| 45 |
|
| 46 |
Skbprio maintains 64 lists (priorities go from 0 to 63). |
| 47 |
When a packet is enqueued, it gets inserted at the |
| 48 |
.B tail |
| 49 |
of its priority list. When a packet needs to be sent out to the network, it is |
| 50 |
taken from the head of the highest priority list. When the queue is full, |
| 51 |
the packet at the tail of the lowest priority list is dropped to serve the |
| 52 |
ingress packet - if it is of higher priority, otherwise the ingress packet is |
| 53 |
dropped. This algorithm allocates as much bandwidth as possible to high |
| 54 |
priority packets, while only servicing low priority packets when |
| 55 |
there is enough bandwidth. |
| 56 |
|
| 57 |
.SH PARAMETERS |
| 58 |
.TP |
| 59 |
limit |
| 60 |
Maximum queue size specified in packets. It defaults to 64. |
| 61 |
The range for this parameter is [0, UINT32_MAX]. |
| 62 |
|
| 63 |
.SH SEE ALSO |
| 64 |
.BR tc-prio (8), |
| 65 |
.BR tc-skbedit (8) |
| 66 |
|
| 67 |
.SH AUTHORS |
| 68 |
Nishanth Devarajan <[email protected]>, Michel Machado <[email protected]> |
| 69 |
|
| 70 |
This manpage maintained by Bert Hubert <[email protected]> |