[Commotion-dev] static BSSID?

Wed May 30 03:09:34 UTC 2012

No problem, happy to help.

Sean

On Tue, May 29, 2012 at 10:34 PM, Josh King <jking at chambana.net> wrote:
> Hi Sean,
>
> I talked with Hans about this last week. Looks great, I think it will
> be a workable approach. I'll work on adding this to the next release of
> CommotionWRT, so we can test it out. Thanks!
>
> On Fri 18 May 2012 05:04:42 PM EDT, Sean McIntyre wrote:
>> Hey folks,
>>
>> Talked about this with Hans today. We came up with the following algorithm:
>>
>> BSSID = MD4(ESSID)[0:4] ++ channel 2-byte integer
>>
>> This is read in English as: "The six byte BSSID is the first four
>> bytes of the MD4 hash of ESSID concatenated with the lowest two bytes
>> of the channel."
>>
>> The rationale for this is that MD4 is uniformly distributed, and the
>> first four bytes of MD4 are presumably uniformly distributed. (Should
>> be distributed enough.) Then add on the bytes of the channel so that
>> same ESSIDs don't collide. MD4 was an arbitrary function choice, and
>> security is not an issue. :)
>>
>> I found a public domain implementation of MD4
>> <http://openwall.info/wiki/people/solar/software/public-domain-source-code/md4>
>> and made a simple get_bssid() function the same as the above algorithm
>> (attached).
>>
>> Signature for get_bssid:
>>
>> get_bssid(char *essid, unsigned int channel, unsigned char *bssid) //
>> (result goes in bssid)
>>
>> Not sure how to add this in, so hope one of you goes ahead and does
>> that, or let me know how I can be of further service.
>>
>> Sean
>>
>> On Thu, May 17, 2012 at 4:23 PM, L. Aaron Kaplan <aaron at lo-res.org> wrote:
>>>
>>> On May 17, 2012, at 10:12 PM, Hans-Christoph Steiner wrote:
>>>
>>>>
>>>> If you have multiple mesh nodes radiating on the same frequency in the same area, so that the radio waves overlap, is there ever a time with an OLSR mesh
>>>> where it would be preferable to have those nodes on separate BSSIDs?  That's the case I'm talking about.
>>>>
>>> Yes, you actually could think about that.
>>> Why? Because you limit the number of neighbors this way. And then you get smaller OLSR  packets.
>>>
>>> But in general: if the density of nodes on the same frequencies is too high, then you are in trouble. No matter which protocol.
>>>
>>>> For Ad-hoc networks, if your neighbor happens to be radiating on the same frequency, that doesn't mean you want to your ad-hoc network to merge with theirs.
>>> Exactly!
>>>
>>>> That would cause confusion with DHCP servers, different IP ranges and devices on the network, different internet connections, etc.
>>>>
>>> ACK.
>>>
>>> Ok, maybe I misunderstood you and was just nit-picking on one particular sentence.
>>>
>>>> .hc
>>>>
>>>> On May 17, 2012, at 11:16 AM, L. Aaron Kaplan wrote:
>>>>
>>>>> Maybe we mean the same thing but it came across differently or I am simply nit-picking now.
>>>>> But as it is written here:
>>>>> "what I can see since more mesh nodes on the same SSID mean a better network"
>>>>>
>>>>> is  not true
>>>>>
>>>>> I would say the more nodes on *different* channels but somehow connected to each other (LAN), the better the network.
>>>>>
>>>>> Use cable whenever you can :) It keeps those darn radio waves contained hehe.
>>>>>
>>>>> I think I have a very very neat solution for this. Can show it to you next time we meet.
>>>>>
>>>>> a.
>>>>>
>>>>> On May 13, 2012, at 4:08 AM, Hans-Christoph Steiner wrote:
>>>>>
>>>>>>
>>>>>> My guess is that this is not a good idea to include in the madwifi driver, since it goes against how the Ad-hoc standard is supposed to work, and it would hinder interoperability with other implementations outside of madwifi (other Linux drivers, Windows, BSD, Mac OS X, etc.)
>>>>>>
>>>>>> But for mesh use, this makes perfect sense from what I can see since more mesh nodes on the same SSID mean a better network, versus with standard Ad-hoc, multiple networks with the same SSID would mean interference, both on the wifi level, and things like multiple DHCP servers, etc.
>>>>>>
>>>>>> .hc
>>>>>>
>>>>>> On May 12, 2012, at 9:57 PM, Josh King wrote:
>>>>>>
>>>>>>> BSSID in an ad-hoc network definitely should be statically set in order
>>>>>>> to avoid network cell partitioning problems. I remember that hashing
>>>>>>> the SSID was a proposed solution for this in the madwifi driver, but I
>>>>>>> don't know if it was ever implemented (I vaguely recall some objection
>>>>>>> to this as a bad idea, but I cannot for the life of me find the
>>>>>>> reference to that argument, or think myself of why it might not be a
>>>>>>> good idea). In any event, there's an old patch posted for the madwifi
>>>>>>> driver that provides some sample code that could provide some guidance
>>>>>>> w/ respect to hashing:
>>>>>>>
>>>>>>> http://permalink.gmane.org/gmane.linux.drivers.madwifi.devel/4787
>>>>>>>
>>>>>>> On Sat 12 May 2012 09:02:40 PM EDT, Jeremy Lakeman wrote:
>>>>>>>> No, we haven't chosen a hash function at all.
>>>>>>>>
>>>>>>>> On Sun, May 13, 2012 at 3:27 AM, Hans-Christoph Steiner
>>>>>>>> <hans at guardianproject.info> wrote:
>>>>>>>>>
>>>>>>>>> That's a really nice, tight idea.  I like it.  Do you have an algorithm for it yet?
>>>>>>>>>
>>>>>>>>> .hc
>>>>>>>>>
>>>>>>>>> On May 11, 2012, at 10:06 PM, Jeremy Lakeman wrote:
>>>>>>>>>
>>>>>>>>>> I propose that the BSSID is generated as a hash of the SSID. This way
>>>>>>>>>> each mesh network will have a unique, but static BSSID without needing
>>>>>>>>>> the user to know it.
>>>>>>>>>>
>>>>>>>>>> On Sat, May 12, 2012 at 3:33 AM, Hans-Christoph Steiner
>>>>>>>>>> <hans at guardianproject.info> wrote:
>>>>>>>>>>>
>>>>>>>>>>> The trickiest thing for the user of the OlsrMeshTether app remains the ad-hoc network associating.  If a bunch of phones are starting up at the same time, they can easily start associating with themselves in clusters.  Then there is multiple meshes that can't talk to each other.
>>>>>>>>>>>
>>>>>>>>>>> It seems a simple solution to this is to just hard-code the BSSID for Commotion in general as a default, then allow people to change this if they really want to.  I propose using CC:CC:CC:CC:CC:CC as the static BSSID.
>>>>>>>>>>>
>>>>>>>>>>> Is there any reason why we wouldn't want to use a static BSSID as the default?
>>>>>>>>>>>
>>>>>>>>>>> .hc
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Commotion-dev mailing list
>>>>>>>>>>> Commotion-dev at lists.chambana.net
>>>>>>>>>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>>>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Commotion-dev mailing list
>>>>>>>> Commotion-dev at lists.chambana.net
>>>>>>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Josh King
>>>>>>>
>>>>>>> "I am an Anarchist not because I believe Anarchism is the final goal,
>>>>>>> but because there is no such thing as a final goal." -Rudolf Rocker
>>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Commotion-dev mailing list
>>>>>> Commotion-dev at lists.chambana.net
>>>>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>>>>>>
>>>>>
>>>>
>>>
>>>
>>> _______________________________________________
>>> Commotion-dev mailing list
>>> Commotion-dev at lists.chambana.net
>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>>>
>>>
>>>
>>> _______________________________________________
>>> Commotion-dev mailing list
>>> Commotion-dev at lists.chambana.net
>>> http://lists.chambana.net/mailman/listinfo/commotion-dev
>
> --
> Josh King
>
> "I am an Anarchist not because I believe Anarchism is the final goal,
> but because there is no such thing as a final goal." -Rudolf Rocker
>