Langue: en

Version: 361857 (ubuntu - 24/10/10)

Section: 4 (Pilotes et protocoles réseau)

BSD mandoc


ng_netflow - Cisco's NetFlow implementation


In sys/types.h In netinet/in.h In netgraph/netflow/ng_netflow.h


The node implements Cisco's NetFlow export protocol on a router running Fx . The node listens for incoming traffic and identifies unique flows in it. Flows are distinguished by endpoint IP addresses, TCP/UDP port numbers, ToS and input interface. Expired flows are exported out of the node in NetFlow version 5 UDP datagrams. Expiration reason can be one of the following:
RST or FIN TCP segment.
Active timeout.
Flows cannot live more than the specified period of time. The default is 1800 seconds (30 minutes).
Inactive timeout.
A flow was inactive for the specified period of time. The default is 15 seconds.

Export information is stored in NetFlow version 5 datagrams.


This node type supports up to NG_NETFLOW_MAXIFACES hooks named iface0 , iface1 etc., and the same number of hooks named out0 , out1 etc., plus a single hook named export By default (ingress NetFlow enabled) node does NetFlow accounting of data received on iface* hooks. If corresponding out hook is connected, unmodified data is bypassed to it, otherwise data is freed. If data is received on out hook, it is bypassed to corresponding iface hook without any processing (egress NetFlow disabled by default). When full export datagram is built it is sent to the export hook. In normal operation, the export hook is connected to the inet/dgram/udp hook of the ng_ksocket4 node.


This node type supports the generic control messages, plus the following:
Returns some node statistics and the current timeout values in a Vt struct ng_netflow_info .
Returns information about the iface N hook. The hook number is passed as an argument.
Sets data link type on the iface N hook. Currently, supported types are raw IP datagrams and Ethernet. This message type uses Vt struct ng_netflow_setdlt as an argument:
 struct ng_netflow_setdlt {
         uint16_t iface;         /* which iface to operate on */
         uint8_t  dlt;           /* DLT_XXX from bpf.h */

The requested iface N hook must already be connected, otherwise message send operation will return an error.

In some cases, may be unable to determine the input interface index of a packet. This can happen if traffic enters the node before it comes to the system interface's input queue. An example of such a setup is capturing a traffic between synchronous data line and ng_iface4. In this case, the input index should be associated with a given hook. The interface's index can be determined via if_nametoindex3 from userland. This message requires Vt struct ng_netflow_setifindex as an argument:
 struct ng_netflow_setifindex {
         u_int16_t iface;        /* which iface to operate on */
         u_int16_t index;        /* new index */

The requested iface N hook must already be connected, otherwise the message send operation will return an error.

Sets values in seconds for NetFlow active/inactive timeouts. This message requires Vt struct ng_netflow_settimeouts as an argument:
 struct ng_netflow_settimeouts {
         uint32_t        inactive_timeout;
         uint32_t        active_timeout;
Sets configuration for the specified interface. This message requires Vt struct ng_netflow_setconfig as an argument:
 struct ng_netflow_setconfig {
         u_int16_t iface;
         u_int32_t conf;
 #define NG_NETFLOW_CONF_INGRESS         1
 #define NG_NETFLOW_CONF_EGRESS          2
 #define NG_NETFLOW_CONF_ONCE            4
 #define NG_NETFLOW_CONF_THISONCE        8

Configuration is a bitmask of several options. Option NG_NETFLOW_CONF_INGRESS enabled by default enables ingress NetFlow generation (for data coming from ifaceX hook). Option NG_NETFLOW_CONF_EGRESS enables egress NetFlow (for data coming from outX hook). Option NG_NETFLOW_CONF_ONCE defines that packet should be accounted only once if it several times passes via netflow node. Option NG_NETFLOW_CONF_THISONCE defines that packet should be accounted only once if it several times passes via exactly this netflow node. Last two options are important to avoid duplicate accounting when both ingress and egress NetFlow are enabled.

This control message asks a node to dump the entire contents of the flow cache. It is called from flowctl(8), not directly from ngctl(8). See also Sx BUGS section.


Most binary control messages have an ASCII equivalent. The supported ASCII commands are:
Qq Li info
Qq Li ifinfo %u
Qq Li setdlt { iface = %u dlt = %u }
Qq Li setifindex { iface = %u index = %u }
Qq Li settimeouts { inactive = %u active = %u }
Qq Li setconfig { iface = %u conf = %u }


This node shuts down upon receipt of a NGM_SHUTDOWN control message, or when all hooks have been disconnected.


The simplest possible configuration is one Ethernet interface, where flow collecting is enabled.
 /usr/sbin/ngctl -f- <<-SEQ
         mkpeer fxp0: netflow lower iface0
         name fxp0:lower netflow
         connect fxp0: netflow: upper out0
         mkpeer netflow: ksocket export inet/dgram/udp
         msg netflow:export connect inet/

This is a more complicated example of a router with 2 NetFlow-enabled interfaces fxp0 and ng0 Note that the ng0: node in this example is connected to ng_tee4. The latter sends us a copy of IP packets, which we analyze and free. On fxp0: we do not use tee, but send packets back to ether node.

 /usr/sbin/ngctl -f- <<-SEQ
         # connect ng0's tee to iface0 hook
         mkpeer ng0:inet netflow right2left iface0
         name ng0:inet.right2left netflow
         # set DLT to raw mode
         msg netflow: setdlt { iface=0 dlt=12 }
         # set interface index (5 in this example)
         msg netflow: setifindex { iface=0 index=5 }
         # Connect fxp0: to iface1 and out1 hook
         connect fxp0: netflow: lower iface1
         connect fxp0: netflow: upper out1
         # Create ksocket node on export hook, and configure it
         # to send exports to proper destination
         mkpeer netflow: ksocket export inet/dgram/udp
         msg netflow:export connect inet/


netgraph(4), ng_ether4, ng_iface4, ng_ksocket4, ng_tee4, flowctl(8), ngctl(8)


An -nosplit The node type was written by An Gleb Smirnoff Aq , based on ng_ipacct written by An Roman V. Palagin Aq .


Cache snapshot obtained via NGM_NETFLOW_SHOW command may lack some percentage of entries under severe load.

The node type does not fill in AS numbers. This is due to the lack of necessary information in the kernel routing table. However, this information can be injected into the kernel from a routing daemon such as GNU Zebra. This functionality may become available in future releases.