"alfred is a user space daemon to efficiently[tm] flood the network with useless data - like vis, weather data, network notes, etc"

• Marek Lindner, 2012

alfred is a user space daemon for distributing arbitrary local information over the mesh/network in a decentralized fashion. This data can be anything which appears to be useful - originally designed to replace the batman-adv visualization (vis), you may distribute hostnames, phone books, administration information, DNS information, the local weather forecast ...

alfred runs as daemon in the background of the system. A user may insert information by using the alfred binary on the command line, or use special programs to communicate with alfred (done via unix sockets). alfred then takes care of distributing the local information to other alfred servers on other nodes. This is done via IPv6 link-local multicast, and does not require any configuration. A user can request data from alfred, and will receive the information available from all alfred servers in the network. Alternatively, alfred can be configured to distribute the local information via IPv4 multicast. This is configured by setting the IPv4 multicast group address in the -4 option.

alfred depends on:

• librt (usually part of libc)
• IPv6 support in the kernel/host system
• libnl-3 - support for netlink sockets
• libnl-3-genl - support for generic netlink messages

and optionally:

• libgps - if you want to distribute GPS information
• libcap - if you want extra security by dropping unneeded privileges

To compile alfred, simply type:

$ make

This will compile alfred, batadv-vis & alfred-gpsd. To install, use:

$ make install

(with the right privileges).

If you don't want to compile batadv-vis, add the directive CONFIG_ALFRED_VIS=n:

$ make CONFIG_ALFRED_VIS=n
$ make CONFIG_ALFRED_VIS=n install

If you don't want to compile alfred-gpsd, add the directive CONFIG_ALFRED_GPSD=n:

$ make CONFIG_ALFRED_GPSD=n
$ make CONFIG_ALFRED_GPSD=n install

If don't want to compile with libcap to drop privileges, use:

$ make CONFIG_ALFRED_CAPABILITIES=n
$ make CONFIG_ALFRED_CAPABILITIES=n install

First, alfred must run as daemon (server) in background to be used. This can either be done by some init-scripts from your distribution (if you have received alfred as a package with your distribution). Please see their documentation how to configure alfred in this case. In any event, you can still run alfred from the command line. The relevant options are (for a full list of options, run alfred -h):

-i, --interface
	specify the interface to listen on
-b	specify the batman-adv interface configured on the system (default: bat0). use 'none' to disable the batman-adv based best server selection
-m, --master	start up the daemon in master mode, which accepts data from slaves and syncs it with other masters

The -b option is optional, and only needed if you run alfred on a batman-adv interface not called bat0, or if you don't use batman-adv at all (use '-b none'). In this case, alfred will still work but will not be able to find the best next master server based on metrics.

alfred servers may either run as master or slave in the network. Masters will announce their status via broadcast, so that slaves can find them. Slaves will then send their data to their nearest master (based on TQ). Masters will exchange their data (which they have received from slaves or got on their own) with other masters. By using masters and slaves, overhead can be reduced while still keeping redundancy (by having multiple masters). Obviously, at least one master must be present in the network to let any data exchange happen. Also having all nodes in master mode is possible (for maximum decentrality and overhead).

To put it together, let us start alfred in master mode on our bridge br0 (assuming that this bridge includes the batman interface bat0):

$ alfred -i br0 -m

Now that the server is running, let us input some data. This can be done by using the alfred binary in client mode from the command line:

$ cat /etc/hostname | alfred -s 64

This will set the hostname as data for datatype 64. Note that 0 - 63 are reserved (please send us an e-mail if you want to register a datatype), and can not be used on the commandline. We skipped the version parameter allowing you to assign a version to your data which can be filtered by other alfred users. Skipping the parameter entirely has the same effect as setting the parameter to 0 ('-V 0').

After the hostname has been set on a few alfred hosts, the can be retrieved again:

$ alfred -r 64
{ "fe:f1:00:00:01:01", "OpenWRT-node-1\x0a" },
{ "fe:f1:00:00:02:01", "OpenWRT-node-2\x0a" },
{ "fe:f1:00:00:03:01", "OpenWRT-node-3\x0a" },

Note that the information must be periodically written again to alfred, otherwise it will timeout and alfred will forget about it (after 10 minutes).

One final remark on terminology: If we talk about "servers" and "clients" in alfred, we mean the local processes on one machine which talk to each other via unix sockets (client connects and talks to servers). On the other hand, "slaves" and "masters" are the roles alfred can take over in the network between different machines (slaves send information to masters).

batadv-vis can be used to visualize your batman-adv mesh network. It read the neighbor information and local client table and distributes this information via alfred in the network. By gathering this local information, any vis node can get the whole picture of the network.

batadv-vis, similar to to alfred, combines server (daemon) and client functionality in the 'batadv-vis' binary. The batadv-vis server must be started to let batadv-vis work:

$ batadv-vis -i bat0 -s

This server will read the neighbor and client information from batman-adv every 10 seconds and set it in alfred via unix socket. Obviously, the alfred server must run too to get this information set.

To get a graphviz-compatible vis output, simply type:

$ batadv-vis
digraph {
        subgraph "cluster_fe:f0:00:00:04:01" {
                "fe:f0:00:00:04:01"
        }
        "fe:f0:00:00:04:01" -> "fe:f0:00:00:05:01" [label="1.000"]
        "fe:f0:00:00:04:01" -> "fe:f0:00:00:03:01" [label="1.004"]
        "fe:f0:00:00:04:01" -> "00:00:43:05:00:04" [label="TT"]
        "fe:f0:00:00:04:01" -> "fe:f1:00:00:04:01" [label="TT"]
        subgraph "cluster_fe:f0:00:00:02:01" {
                "fe:f0:00:00:02:01"
        }
        "fe:f0:00:00:02:01" -> "fe:f0:00:00:03:01" [label="1.000"]
        "fe:f0:00:00:02:01" -> "fe:f0:00:00:01:01" [label="1.008"]
        "fe:f0:00:00:02:01" -> "fe:f0:00:00:08:01" [label="1.000"]
        "fe:f0:00:00:02:01" -> "fe:f1:00:00:02:01" [label="TT"]
        "fe:f0:00:00:02:01" -> "00:00:43:05:00:02" [label="TT"]
        subgraph "cluster_fe:f0:00:00:08:01" {
                "fe:f0:00:00:08:01"
        }
[...]
}

For a json line formatted output, use:

$ batadv-vis -f json
{ "primary" : "fe:f0:00:00:04:01" }
{ "router" : "fe:f0:00:00:04:01", "neighbor" : "fe:f0:00:00:05:01", "label" : "1.000" }
{ "router" : "fe:f0:00:00:04:01", "neighbor" : "fe:f0:00:00:03:01", "label" : "1.008" }
{ "router" : "fe:f0:00:00:04:01", "gateway" : "00:00:43:05:00:04", "label" : "TT" }
{ "router" : "fe:f0:00:00:04:01", "gateway" : "fe:f1:00:00:04:01", "label" : "TT" }
{ "primary" : "fe:f0:00:00:02:01" }
{ "router" : "fe:f0:00:00:02:01", "neighbor" : "fe:f0:00:00:03:01", "label" : "1.000" }
{ "router" : "fe:f0:00:00:02:01", "neighbor" : "fe:f0:00:00:01:01", "label" : "1.016" }
{ "router" : "fe:f0:00:00:02:01", "neighbor" : "fe:f0:00:00:08:01", "label" : "1.000" }
{ "router" : "fe:f0:00:00:02:01", "gateway" : "fe:f1:00:00:02:01", "label" : "TT" }
{ "router" : "fe:f0:00:00:02:01", "gateway" : "00:00:43:05:00:02", "label" : "TT" }
{ "primary" : "fe:f0:00:00:08:01" }
[...]

and for output where the complete document is json, use:

$ batadv-vis -f jsondoc
{
  "source_version" : "2013.3.0-14-gcd34783",
  "algorithm" : 4,
  "vis" : [
    { "primary" : "fe:f0:00:00:04:01",
      "neighbors" : [
         { "router" : "fe:f0:00:00:04:01",
           "neighbor" : "fe:f0:00:00:05:01",
           "metric" : "1.000" },
         { "router" : "fe:f0:00:00:04:01",
           "neighbor" : "fe:f0:00:00:03:01",
           "metric" : "1.008" }
      ],
      "clients" : [
         "00:00:43:05:00:04",
         "fe:f1:00:00:04:01"
      ]
    },
    { "primary" : "fe:f0:00:00:02:01",
      "neighbors" : [
         { "router" : "fe:f0:00:00:02:01",
           "neighbor" : "fe:f0:00:00:03:01",
           "metric" : "1.000" },
         { "router" : "fe:f0:00:00:02:01",
           "neighbor" : "fe:f0:00:00:01:01",
           "metric" : "1.016" },
         { "router" : "fe:f0:00:00:02:01",
           "neighbor" : "fe:f0:00:00:08:01",
           "metric" : "1.000" }
      ],
      "clients" : [
        "fe:f1:00:00:02:01",
        "00:00:43:05:00:02"
      ]
    },
    { "primary" : "fe:f0:00:00:08:01",
[...]

Alfred-gpsd can be used to distibute GPS location information about your batman-adv mesh network. This information could be, for example, combined with Vis to visualize your mesh topology with true geographic layout. For mobile or nomadic nodes, Alfred-gpsd, can get location information from gpsd. Alternatively, a static location can be passed on the command line, which is useful for static nodes without a GPS.

Alfred-gpsd, similar to to alfred, combines server (daemon) and client functionality in the 'alfred-gpsd' binary. The alfred-gpsd server must be started to distribute location information. When retrieving location information from gpsd, it should be started with:

$ alfred-gpsd -s

For a static location, use:

$ alfred-gpsd -s -l 48.858222,2.2945,358

This server will set the location in alfred via unix socket. Obviously, the alfred server must run too to get this information set. When using gpsd, it updates alfred every 2 seconds. With a static location, the update it made every 5 minutes.

To get JSON formatted output, use:

$ alfred-gpsd
[
  { "source" : "f6:00:48:13:d3:1e", "tpv" : {"class":"TPV","tag":"RMC","device":"/dev/ttyACM0","mode":3,"time":"2013-10-01T10:43:20.000Z","ept":0.005,"lat":52.575485000,"lon":-1.339716667,"alt":122.500,"epx":10.199,"epy":15.720,"epv":31.050,"track":0.0000,"speed":0.010,"climb":0.000,"eps":31.44} },
  { "source" : "8e:4c:77:b3:65:b4", "tpv" : {"class":"TPV","device":"command line","time":"2013-10-01T10:43:05.129Z","lat":48.858222,"lon":2.2945,"alt":358.000000,"mode":3} }
]

See gpsd_json(5) for documentation of the tpv object.

Alfred currently requires special capabilities and access rights to work correctly. The user root is normally the only user having these capabilities/rights on a standard Linux system.

Operations requiring special capabilities:

• bind to device
• creating the unix socket
• accessing the debugfs filesystem

The first operation can still be executed when the admin grants the special capability CAP_NET_RAW+CAP_NET_ADMIN to anyone executing the alfred binary. The unix socket can also be moved using the parameter '-u' to a different directory which can be accessed by the user:

$ sudo setcap cap_net_admin,cap_net_raw+ep alfred
$ ./alfred -u alfred.sock -i eth0

The user running alfred must still be in a group which is allowed to access /sys/kernel/debugfs to correctly choose best neighbors for communication. It is possible (but not recommended) to disable the neighbor selection/prioritization using the parameter '-b none'.

alfred, batadv-vis and alfred-gpsd are licensed under the terms of version 2 of the GNU General Public License (GPL). Please see the LICENSE file.

The file "packet.h" is an exception and not licensed with the GPL. Instead, it is licensed using ISC license (see the head of this file). This allows programs to include this header file (e.g. for communicating with alfred via unix sockets) without enforcing the restrions of the GPL license on this third party program.

As alfred was developed to help on batman-adv, we share communication channels. Please send us comments, experiences, questions, anything :)

IRC:

#batman on irc.freenode.org

Mailing-list:

[email protected] (optional subscription at https://lists.open-mesh.org/mm/listinfo/b.a.t.m.a.n)

If you have test reports/patches/ideas, please read the wiki for further instruction on how to contribute:

https://www.open-mesh.org/projects/open-mesh/wiki/Contribute

You can also contact the Authors:

• Marek Lindner <[email protected]>
• Simon Wunderlich <[email protected]>