forked from ntop/PF_RING
-
Notifications
You must be signed in to change notification settings - Fork 0
/
README.FIRST
93 lines (77 loc) · 3.5 KB
/
README.FIRST
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
PF_RING is a Linux kernel module and user-space framework that allows
you to process packets at high-rates while providing you a consistent
API for packet processing applications. This directory contains:
drivers PF_RING optimized drivers
drivers/ PF_RING-aware/ZC drivers (suggested option)
userland/ User space code
userland/lib/ User space library
userland/libpcap-XXX-ring/ Libpcap enhanced with PF_RING support
userland/examples/ PF_RING sample applications
userland/examples_zc/ PF_RING ZC sample applications
userland/snort/ Snort DAQ module for using snort over PF_RING
PF_RING Drivers Models
----------------------
1. PF_RING-aware/ZC drivers
These drivers are called "aware" as they are optimized for PF_RING. They
can be used in two ways. As standard drivers (packets continue to pass
inside the Linux stack), example:
pfcount -i eth1
or in zero copy, for completely bypassing the Linux stack in both RX and
TX, and thus achieve line rate. As soon as one application opens the
interface in zero copy (ZC), such interface (while still visible by
ifconfig) cannot longer be used for standard networking (e.g. ping or SSH);
this until the interface is open in ZC, and once it is closed it returns
to be immediately available for standard Linux networking. In order to open
it in ZC you must prepend 'zc:' to the interface name, example:
pfcount -i zc:eth1
2. ZC drivers
These drivers have been used to implement kernel bypass until PF_RING 5.x
and since the ZC introduction they will be available for some more time
but they are deprecated in favour of ZC that offer the same performance
while featuring much better flexibility. In future PF_RING releases, we
expect to discontinue thes drivers.
PF_RING from Virtual Machines (KVM)
-----------------------------------
With the introduction of PF_RING ZC, you can finally exploit acceleration
even from withing a KVM virtual machine without using techniques such as
PCIe bypass. This means that a VM can dynamically open network interfaces
in ZC, implement complex packet processing topologies, all at 10 Gbit
line rate. Please refer to userland/examples_zc/README.kvm for more
information on how to exploit PF_RING ZC from KVM.
As future activity, we might evaluate to port it to other hypervisors,
but at the moment we target only KVM.
Compilation
-----------
You need to have installed the basic compilation tools and libraries.
On Ubuntu you can do
# apt-get install build-essential bison flex linux-headers-$(uname -r) libnuma-dev
At this point you can compile the module
# make
Installation
------------
# sudo su
# cd kernel; make install
# cd ../userland/lib; make install
Testing PF_RING
---------------
In PF_RING/userland/examples we have coded several applications you can use
for testing PF_RING. In particular
- pfcount allows you to capture packets
- pfsend to replay (or generate) traffic
You can look at the application source code for learning how you can use
the PF_RING API.
Remember that before running any appliction, you need to load the pf_ring.ko
kernel modules:
# sudo su
# insmod ./kernel/pf_ring.ko
Documentation
-------------
If you need an extensive documentation, you can refer to the manual that
is stored in the doc/ directory. In such directory you will also find
individual README files for the various network adapters supported by
PF_RING.
The API documentation is available via doxygen, running
# make documentation
it will be generated in doc/html.
-------------------------------------
(C) 2004-15 - ntop.org