[B.A.T.M.A.N.] Maintainer of vis and batmand

[B.A.T.M.A.N.] Maintainer of vis and batmand

[elektra]

Hi Sven -

AFAIK *vis* has no maintainer at the moment.
I am currently working on the routing loop issue in batmand and link-local OGM aggregation, in order to improve scalability. 
Will do some testing at the Battlemesh. 

Cheers,
Elektra

Re: [B.A.T.M.A.N.] Maintainer of vis and batmand

[Marek Lindner]

Hi Elektra,

> AFAIK *vis* has no maintainer at the moment.
> I am currently working on the routing loop issue in batmand and link-local
> OGM aggregation, in order to improve scalability. Will do some testing at
> the Battlemesh.

meaning you are maintaining batmand ? Should we change our wiki pages that 
claim batmand has no maintainer ?

Regards,
Marek

Re: [B.A.T.M.A.N.] Maintainer of vis and batmand

[elektra]

Hi Marek -

> meaning you are maintaining batmand ? Should we change our wiki pages that 
> claim batmand has no maintainer ?

yes, I am glad to do this. I feel there is room for a simple Layer 3 implementation of Batman beside the exiting batman-adv. And hey, I have started the Batman project with the early bits of this code together with Thomas. Call it nostalgia ;-) And I am also thinking about using a scaled-down batmand version as a proactive mesh protocol for microcontrollers and for shortwave radio meshes with low bandwidth.

As a sidenote:

I have been playing with batman-adv on the Mesh-Potato, and I will come up with a batman-adv based version of the MP firmware. It opens up a exciting set of new features. Bridging a VAP AP with the LAN port and the bat0 interface is really cool.

However I have seen interesting new challenges. For example, if the MPs just get their IP configuration from a DHCP server this eliminates the process of configuring an IP for each device. That is cool. On the other hand, not knowing which IP is assigned to which device introduces new problems. It makes debugging radio problems and hardware issues much harder in the field. So picking up a phone and punching in a static IP via the keypad for each device is not all bad, if you just want to be able to call one MP from the other by dialling the last octet of its IP. 

Is there a special way of handling IPv4 ARP packets in batman-adv? I figure ARP must be much more robust in a Layer 3 mesh. 
 
Cheers,
Elektra

Re: [B.A.T.M.A.N.] Maintainer of vis and batmand

[Marek Lindner]

Hi,

> I have been playing with batman-adv on the Mesh-Potato, and I will come up
> with a batman-adv based version of the MP firmware.

cool!


> However I have seen interesting new challenges. For example, if the MPs
> just get their IP configuration from a DHCP server this eliminates the
> process of configuring an IP for each device. That is cool. On the other
> hand, not knowing which IP is assigned to which device introduces new
> problems. It makes debugging radio problems and hardware issues much
> harder in the field. So picking up a phone and punching in a static IP via
> the keypad for each device is not all bad, if you just want to be able to
> call one MP from the other by dialling the last octet of its IP.

I don't really understand where you see the "new" challenges. If you require 
static IP addresses you can deal with that in the same way you did without 
batman-adv. You could configure alias interfaces, calculate an IP out of the 
MAC or even use DHCP to assign always the same IP address to a certain client.
How your setup looks like very much depends on your requirements.


> Is there a special way of handling IPv4 ARP packets in batman-adv? I figure
> ARP must be much more robust in a Layer 3 mesh.

There is no special handling yet (expect for the ARQ mechanism for all 
broadcast packets). I also suspect that ARP requires some more attention but 
nobody has done that so far.

Cheers,
Marek

Re: [B.A.T.M.A.N.] Maintainer of vis and batmand

[Andrew Lunn]

> However I have seen interesting new challenges. For example, if the
> MPs just get their IP configuration from a DHCP server this
> eliminates the process of configuring an IP for each device. That is
> cool. On the other hand, not knowing which IP is assigned to which
> device introduces new problems. It makes debugging radio problems
> and hardware issues much harder in the field. 

Hi Elektra

An option would be to move to IPv6 and use stateless address
autoconfiguration, where the MAC address is used to form part of the
IPv6 address. You can then easily see from the IP address which device
it is, assuming you know its MAC address.

   Andrew

Re: [B.A.T.M.A.N.] Maintainer of vis and batmand

[elektra]

Hi all -

> An option would be to move to IPv6 and use stateless address
> autoconfiguration, where the MAC address is used to form part of the
> IPv6 address. You can then easily see from the IP address which device
> it is, assuming you know its MAC address.

the background of my thoughts was how far we can get with non-technical people just taking a 
bunch of Mesh-Potatoes out of the box, connecting one of them to an existing LAN segment with 
DHCP and Internet uplink. The setup should allow them to make calls amongst each other and 
simplifying the gateway setup, so they can use the Internet and make international calls. The 
MPs should also have a Accesspoint VAP.

This is an attempt of squaring the circle, of course...

Having said that batman-adv is pretty powerful with this regard. I just wanted to initiate some 
public brainstorming about how to solve the problem of identifying each device for debugging 
purposes when using DHCP.

Writing down each devices MAC address and using IPv6 is a feasible approach, but I figure it 
would be not intuitive.

However Mareks suggestion of adding a alias IP (we have a Asterisk IVR extension that can do 
this) is a valid suggestion.  

The requirement to enter a IP via the phone keypad is not exactly zero configuration, but given 
the enormous set of features that we can offer, it is pretty close. Thanks!


Cheers,
Elektra