[Olsr-dev] Bug Secure Plugin - Endianness ?

[Roar Bjørgum Rotvik]

On 06.05.2011 10:14, Henning Rogge wrote:
> On Fri May 6 2011 10:05:13 ZioPRoTo (Saverio Proto) wrote:
>> I've been reported by an "advanced-user" that the secure plugin has
>> problems when mixing in the network nodes with different architectures
>> (example X86 and Ubiquities).
>>
>> His setup works perfectly, until he adds to the network a node with a
>> different arch. Maybe there is some endianess problem in the code ?
> Yes, might be... I don't think anyone has ever used this plugin.
>
> I think the design of the plugin (authentication of routing packages based on
> shared group key and timestamps) make it broken by design.

Hi,

I'm one of the guys that thought of and designed this secure plug in.
It was designed for use in a controlled net, i.e. all the allowed nodes 
has a preshared key obtained out-of-band (i.e. not part of olsr or 
network traffic) that kept all non authorised nodes away from 
interfering with olsr traffic.

It was mostly a proof of concept for a "small" network (not 1000 of 
nodes) where we wanted a basic level of routing security. It is not "the 
final" security solution for authentication in the network and should 
not be used for that. But it should be enough for small and confined 
networks.

What do you believe is "broken" in it's design? Remember that the design 
criteria was simple, small and low overhead (hence no PKI).

Regards,
Roar Bjørgum Rotvik