Hello,
based on the motivation that IoT devices do not have many entropy
sources available, and also on the fact that trusting in a single
source may be risky, I have researched some additional ways to collect
randomness for the device, using bluetooth technology for that
purpose.
For those who are interested in:
Github: https://github.com/AAGiron/Bluerandom
Presentation: https://www.youtube.com/watch?v=iAKaKAmqCFo
I also want to say thanks in advance to Emil Lenngren for his valuable
comments, here in this mailing list (back in 2019), and also to the
Bluez community.
--
Alexandre Augusto Giron
Ola Alexandre,
Que coincidência ver alguem da UTFPR por aqui, fiz meu mestrado ai em
2003, muito legal o seu trabalho. So por curiosidade, voce chegou a
usar o BT_HCI_CMD_LE_RAND como fonte de entropia? No Zephyr
(https://www.zephyrproject.org/) agente usa bastante porque o
controlador pode usar o sinal de interferência direto da camada do
radio, abaixo da interface HCI, e segundo o pessoal da Nordic tem uma
qualidade bem maior do que por exemplo RSSI porque nao depende da
quantidade de dispositivos em volta.
On Thu, Oct 15, 2020 at 2:59 PM Alexandre Augusto
<[email protected]> wrote:
>
> Hello,
>
> based on the motivation that IoT devices do not have many entropy
> sources available, and also on the fact that trusting in a single
> source may be risky, I have researched some additional ways to collect
> randomness for the device, using bluetooth technology for that
> purpose.
>
> For those who are interested in:
>
> Github: https://github.com/AAGiron/Bluerandom
> Presentation: https://www.youtube.com/watch?v=iAKaKAmqCFo
>
> I also want to say thanks in advance to Emil Lenngren for his valuable
> comments, here in this mailing list (back in 2019), and also to the
> Bluez community.
>
> --
> Alexandre Augusto Giron
--
Luiz Augusto von Dentz
Hi Alexandre,
> based on the motivation that IoT devices do not have many entropy
> sources available, and also on the fact that trusting in a single
> source may be risky, I have researched some additional ways to collect
> randomness for the device, using bluetooth technology for that
> purpose.
>
> For those who are interested in:
>
> Github: https://github.com/AAGiron/Bluerandom
> Presentation: https://www.youtube.com/watch?v=iAKaKAmqCFo
>
> I also want to say thanks in advance to Emil Lenngren for his valuable
> comments, here in this mailing list (back in 2019), and also to the
> Bluez community.
I would integrate as kernel driver. The Bluetooth stack could just use its
own LE random generator and an initial passive scan to gather RSSI values.
No need to push this into user space.
Regards
Marcel
Hi Marcel,
excellent. Let me know if any clarification is needed.
Best regards
Em ter., 20 de out. de 2020 às 15:38, Marcel Holtmann
<[email protected]> escreveu:
>
> Hi Alexandre,
>
> > based on the motivation that IoT devices do not have many entropy
> > sources available, and also on the fact that trusting in a single
> > source may be risky, I have researched some additional ways to collect
> > randomness for the device, using bluetooth technology for that
> > purpose.
> >
> > For those who are interested in:
> >
> > Github: https://github.com/AAGiron/Bluerandom
> > Presentation: https://www.youtube.com/watch?v=iAKaKAmqCFo
> >
> > I also want to say thanks in advance to Emil Lenngren for his valuable
> > comments, here in this mailing list (back in 2019), and also to the
> > Bluez community.
>
> I would integrate as kernel driver. The Bluetooth stack could just use its
> own LE random generator and an initial passive scan to gather RSSI values.
> No need to push this into user space.
>
> Regards
>
> Marcel
>
--
Alexandre Augusto Giron
Professor da UTFPR
Doutorando em Ciência da Computação (UFSC)
Sorry, I think I misunderstood your email Marcel, apologies for that.
But yes, I have a sketch for this as a kernel driver (but still non
functional).
Nevertheless, in my repo there is a proof-of-concept version that I
used in the experiments presented in the video.
Thanks for the feedback.
Em ter., 20 de out. de 2020 às 15:38, Marcel Holtmann
<[email protected]> escreveu:
>
> Hi Alexandre,
>
> > based on the motivation that IoT devices do not have many entropy
> > sources available, and also on the fact that trusting in a single
> > source may be risky, I have researched some additional ways to collect
> > randomness for the device, using bluetooth technology for that
> > purpose.
> >
> > For those who are interested in:
> >
> > Github: https://github.com/AAGiron/Bluerandom
> > Presentation: https://www.youtube.com/watch?v=iAKaKAmqCFo
> >
> > I also want to say thanks in advance to Emil Lenngren for his valuable
> > comments, here in this mailing list (back in 2019), and also to the
> > Bluez community.
>
> I would integrate as kernel driver. The Bluetooth stack could just use its
> own LE random generator and an initial passive scan to gather RSSI values.
> No need to push this into user space.
>
> Regards
>
> Marcel
>
--
Alexandre Augusto Giron
Professor da UTFPR
Doutorando em Ciência da Computação (UFSC)