NAME¶
babeld - ad-hoc network routing daemon
SYNOPSIS¶
babeld option... [
-- ]
interface...
DESCRIPTION¶
Babel is a loop-avoiding distance-vector routing protocol roughly based on DSDV
and AODV, but with provisions for link cost estimation and redistribution of
routes from other routing protocols.
While it is optimised for wireless mesh networks, Babel will also work
efficiently on wired networks.
OPTIONS¶
- -m multicast-address
- Specify the link-local multicast address to be used by the protocol. The
default is ff02:0:0:0:0:0:1:6.
- -p port
- Specify the UDP port number to be used by the protocol. The default is
6696.
- -S state-file
- Set the name of the file used for preserving long-term information between
invocations of the babeld daemon. If this file is deleted, the
daemon will run in passive mode for 3 minutes when it is next started (see
-P below), and other hosts might initially ignore it. The default
is /var/lib/babel-state.
- -h hello-interval
- Specify the interval in seconds at which scheduled hello packets are sent
on wireless interfaces. The default is 4 seconds.
- -H wired-hello-interval
- Specify the interval in seconds at which scheduled hello packets are sent
on wired interfaces. The default is 4 seconds.
- -z kind
[,factor]
- Enable diversity-sensitive routing. The value kind defines the
diversity algorithm used, and can be one of 0 (no diversity),
1 (per-interface diversity with no memory), 2 (per-channel
diversity with no memory), or 3 (per-channel diversity with
memory). The value factor specifies by how much the cost of
non-interfering routes is multiplied, in units of 1/256; the default is
128 (i.e. division by 2).
- -M half-time
- Specify the half-time in seconds of the exponential decay used for
smoothing metrics for performing route selection; the value 0 disables
smoothing. The default is 4s.
- -k priority
- Specify the priority value used when installing routes into the kernel.
The default is 0.
- -A priority
- Allow duplicating external routes when their kernel priority is at least
priority. Do not use this option unless you know what you are
doing, as it can cause persistent route flapping.
- -l
- Use IFF_RUNNING (carrier sense) when determining interface
availability.
- -w
- Don't optimise wired links, assume all interfaces are wireless unless
explicitly overridden in the configuration file.
- -s
- Do not perform split-horizon processing on wired interfaces. Split-horizon
is not performed on wireless interfaces.
- -r
- Use a random router-id. The default is to use persistent router-ids
derived from the MAC address of the first interface, which is easier to
debug and more reliably prevents routing loops but may sometimes cause a
node to be unreachable for 120 seconds just after boot.
- -u
- Do not flush unfeasible (useless) routes. This is useful in order to
announce more information to a front-end (see -g).
- -d level
- Debug level. A value of 1 requests a routing table dump at every iteration
through the daemon's main loop. A value of 2 additionally requests tracing
every message sent or received. A value of 3 additionally dumps all
interactions with the OS kernel. The default is 0.
- -g port
- Listen for connections from a front-end on port port.
- -t table
- Use the given kernel routing table for routes inserted by
babeld.
- -T table
- Export routes from the given kernel routing table. This can be specified
multiple times in order to export routes from more than one table.
- -c filename
- Specify the name of the configuration file. This flag can be repeated
multiple times. The default is /etc/babeld.conf.
- -C statement
- Specify a configuration statement directly on the command line.
- -D
- Daemonise at startup.
- -L logfile
- Specify a file to log random ``how do you do?'' messages to. This defaults
to standard error if not daemonising, and to /var/log/babeld.log
otherwise.
- -I pidfile
- Specify a file to write our process id to. The default is
/var/run/babeld.pid.
- interface...
- The list of interfaces on which the protocol should operate.
The configuration file is a sequence of lines each of which specifies a global
option, an interface specification or a filtering rule. Comments are
introduced by an octothorp ``
#'' and terminate at the end of the line.
Global options¶
- protocol-group group
- This specifies the link-local multicast address to be used by the
protocol, and is equivalent to the command-line option -m.
- protocol-port port
- This specifies the UDP port number to be used by the protocol, and is
equivalent to the command-line option -p.
- kernel-priority priority
- This specifies the priority value used when installing routes into the
kernel, and is equivalent to the command-line option -k.
- allow-duplicates priority
- This allows duplicating external routes when their kernel priority is at
least priority. Do not use this option unless you know what you are
doing, as it can cause persistent route flapping.
- keep-unfeasible {true|false}
- This specifies whether to keep unfeasible (useless) routes, and is
equivalent to the command-line option -u.
- random-id {true|false}
- This specifies whether to use a random router-id, and is equivalent to the
command-line option -r.
- debug level
- This specifies the debugging level, and is equivalent to the command-line
option -d.
- local-port port
- This specifies the TCP port on which babeld will listen for
connections from a front-end, and is equivalent to the command-line option
-g.
- export-table table
- This specifies the kernel routing table to use for routes inserted by
babeld, and is equivalent to the command-line option
-t.
- import-table table
- This specifies a kernel routing table from which routes are redistributed
by babeld, and can be specified multiple times with a cumulative
effect. This is equivalent to the command-line option -T.
- link-detect {true|false}
- This specifies whether to use carrier sense for determining interface
availability, and is equivalent to the command-line option -l.
- diversity
{true|false|kind}
- This specifies the diversity algorithm to use; true is equivalent
to kind 3. The default is false (do not use any diversity
algorithm).
- diversity-factor factor
- This specifies by how much the cost of non-interfering routes should be
multiplied, in units of 1/256. The default is 128 (division by 2).
- smoothing-half-life seconds
- This specifies the half-life in seconds of the exponential decay used for
smoothing metrics for performing route selection, and is equivalent to the
command-line option -M.
- deamonise {true|false}
- This specifies whether to daemonize at startup, and is equivalent to the
command-line option -D.
- state-file filename
- This specifies the name of the file used for preserving long-term
information between invocations of the babeld daemon, and is
equivalent to the command-line option -S.
- log-file filename
- This specifies the name of the file used to log random messages to, and is
equivalent to the command-line option -L.
- pid-file filename
- This specifies the name of the file to which babeld writes out its
process id, and is equivalent to the command-line option -I.
Interface configuration¶
An interface is configured by a line with the following format:
- interface name [parameter...]
where
name is the name of the interface (something like
eth0). The
default value of an interface parameter can be specified changed by a line of
the form
- default [parameter...]
Each
parameter can be one of:
- wired {true|false|auto}
- This specifies whether to enable optimisations specific to wired
interfaces. By default, this is determined automatically unless the
-w command-line flag was specified.
- link-quality {true|false|auto}
- This specifies whether link quality estimation should be performed on this
interface. The default is to perform link quality estimation on wireless
interfaces but not on wired interfaces.
- split-horizon {true|false|auto}
- This specifies whether to perform split-horizon processing on this
interface. The default is to never perform split-horizon processing on
wireless interfaces; on wired interfaces, the default depends on the
-s flag.
- rxcost cost
- This defines the cost of receiving frames on the given interface under
ideal conditions (no packet loss); how this relates to the actual cost
used for computing metrics of routes going through this interface depends
on whether link quality estimation is being done. The default is 96 for
wired interfaces, and 256 for wireless ones.
- channel channel
- Sets the channel for this interface. The value channel can be
either an integer, or one of the strings interfering or
noninterfering. The default is to autodetect the channel number for
wireless interfaces, and noninterfering for wired interfaces.
- faraway {true|false}
- This specifies whether the network is "far away", in the sense
that networks behind it don't interfere with networks in front of it. By
default, networks are not far away.
- hello-interval interval
- This defines the interval between hello packets sent on this interface.
The default is specified with the -h and -H command-line
flags.
- update-interval interval
- This defines the interval between full routing table dumps sent on this
interface; since Babel uses triggered updates and doesn't count to
infinity, this can be set to a fairly large value, unless significant
packet loss is expected. The default is four times the hello
interval.
- enable-timestamps {true|false}
- Enable sending timestamps with each Hello and IHU message in order to
compute RTT values. The default is true if max-rtt-penalty
is non-zero (see below), and false otherwise.
- rtt-decay decay
- This specifies the decay factor for the exponential moving average of RTT
samples, in units of 1/256. Must be between 1 and 256, inclusive. Higher
values discard old samples faster. The default is 42.
- rtt-min rtt
- This specifies the minimum RTT, in milliseconds, starting from which we
increase the cost to a neighbour. The additional cost is linear in (rtt -
rtt-min). The default is 10 ms.
- rtt-max rtt
- This specifies the maximum RTT, in milliseconds, above which we don't
increase the cost to a neighbour. The default is 120 ms.
- max-rtt-penalty cost
- This specifies the maximum cost added to a neighbour because of RTT, i.e.
when the RTT is higher or equal than rtt-max. The default is
0, which effectively disables the use of a RTT-based cost.
Filtering rules¶
A filtering rule is defined by a single line with the following format:
- filter selector... action
Filter specifies the filter to which this entry will be added, and can be
one of
in,
out, or
redistribute.
Each
selector specifies the conditions under which the given statement
matches. It can be one of
- ip prefix
- This entry only applies to routes in the given prefix.
- eq plen
- This entry only applies to routes with a prefix length equal to
plen.
- le plen
- This entry only applies to routes with a prefix length less or equal to
plen.
- ge plen
- This entry only applies to routes with a prefix length greater or equal to
plen.
- neigh address
- This entry only applies to routes learned from a neighbour with link-local
address address.
- id id
- This entry only applies to routes originated by a router with router-id
id.
- proto p
- This entry only applies to kernel routes with kernel protocol number
p. If neither proto nor local is specified, this
entry applies to all non-local kernel routes with a protocol different
from "boot".
- local
- This entry only applies to local addresses.
- if interface
- For an input filter, this specifies the interface over which the route is
learned. For an output filter, this specifies the interface over which
this route is advertised. For a redistribute statement, this specifies the
interface over which the route forwards packets.
Action specifies the action to be taken when this entry matches. It can
have one of the following values:
- allow
- Allow this route, without changing its metric (or setting its metric to 0
in case of a redistribute filter).
- deny
- Ignore this route.
- metric value
- For an input or output filter, allow this route after increasing its
metric by value. For a redistribute filter, redistribute this route
with metric value.
If
action is not specified, it defaults to
allow.
By default,
babeld redistributes all local addresses, and no other
routes. In order to make sure that only the routes you specify are
redistributed, you should include the line
- redistribute local deny
as the last line in your configuration file.
EXAMPLES¶
You can participate in a Babel network by simply running
- # babeld wlan0
where
wlan0 is the name of your wireless interface.
In order to gateway between multiple interfaces, just list them all on the
command line:
- # babeld wlan0 eth0 sit1
On an access point, you'll probably want to redistribute some external routes
into Babel:
- # babeld \
-C 'redistribute metric 256' \
wlan0
or, if you want to constrain the routes that you redistribute,
- # babeld \
-C 'redistribute proto 11 ip ::/0 le 64 metric 256' \
-C 'redistribute proto 11 ip 0.0.0.0/0 le 24 metric 256' \
wlan0
FILES¶
- /etc/babeld.conf
- The default location of the configuration file.
- /var/lib/babel-state
- The default location of the file storing long-term state.
- /var/run/babeld.pid
- The default location of the pid file.
- /var/log/babeld.log
- The default location of the log file.
SIGNALS¶
- SIGUSR1
- Dump Babel's routing tables to standard output or to the log file.
- SIGUSR2
- Check interfaces and kernel routes right now, then reopen the log
file.
SECURITY¶
Babel is a completely insecure protocol: any attacker able to inject IP packets
with a link-local source address can disrupt the protocol's operation. This is
no different from unsecured neighbour discovery or ARP.
Since Babel uses link-local IPv6 packets only, there is no need to update
firewalls to allow forwarding of Babel protocol packets. If local filtering is
being done, UDP datagrams to the port used by the protocol should be allowed.
As Babel uses unicast packets in some cases, it is not enough to just allow
packets destined to Babel's multicast address.
BUGS¶
Plenty. This is experimental software, run at your own risk.
SEE ALSO¶
routed(8),
route6d(8),
zebra(8),
ahcpd(8).
AUTHOR¶
Juliusz Chroboczek.