Dear mr Watteyne

I'll explain why the question is relevant for me.


I am researching the possibility whether I can use the tight synchronisation of 
TSCH in a network of sensors/nodes to detect the presence of malware for 
research for a master thesis.

What kind of malware it is, doesn't matter for now. But basically, if malware 
is placed on a node and this action causes clock drift, it might be possible to 
detect this due to the desynchronisation that follows.

In the coming days, I am going to try and test this after getting more familiar 
with OpenWSN and 6TiSCH.
If this is not the case, I could always enforce it by making the node to 
resynchronize everytime new software is added.

Although this scenario would be the best and most fun for me to research, I can 
also try out different aspects. For example: Using TSCH I might be able to 
detect the presence of malware due to a difference in time of a certain process 
or the duration of a packet on the network. I could profile a node and detect 
difference in it between 2 runs of the profiles.
So in conclusion, I would like to do/research intrusion/malware detection in 
IoT using TSCH!
Indeed, it isn't specifically related to 6TiSCH but a possibility to use it 
with/for 6TiSCH. And so I was wondering if I could find some information by 
contacting you.

I don't think a phone call is necessary. Although this is quite important for 
me, I can wait for answers and I can continue on my own if this doesn't work. I 
have multiple options to explore. But thanks for the offer!


Also thanks for the replies so far!

With kind regards


Gregory Braekevelt ?
?

________________________________
Van: Thomas Watteyne <[email protected]>
Verzonden: woensdag 16 november 2016 4:09
Aan: Michael Richardson
CC: [email protected]; Gregory Braekevelt
Onderwerp: Re: [6tisch] Malware in IoT/WSN with 6tisch + Clock Drift

Gregory,

I agree with Michael's response, but I fail to understand the scope/relevance 
of what you are describing.

The firmware running on the mote implements the IEEE802.15.4 TSCH state 
machine, which includes everything related to timing.
If your "malware" is a complete new firmware, it can do anything it wants, 
including introduce an arbitrary drift, blink LEDs, or make coffee.
If your "malware" is a new application running on top of a stack, your software 
architecture should rely on the kernel to prevent this application from 
breaking/influencing the stack. This can be done by having an MPU detect when 
the app accesses memory locations it isnt support to, and killing it.

I feel that this discussion is related to software architecture, not 
protocol/6TiSCH.

Thomas

On Sat, Nov 12, 2016 at 11:49 AM, Michael Richardson 
<[email protected]<mailto:[email protected]>> wrote:

Gregory Braekevelt 
<[email protected]<mailto:[email protected]>>
 wrote:
    > I'm mainly searching for information that proofs or counters the fact
    > that malware (can) cause clock drift.

It could be the case, and for a specific device, and a specific piece of
malware, it could be the case.  I see no reason for there to be general rule
like this.

So, yes, malware could cause clock drift, but I see no reason why malware
MUST cause clock drift.


_______________________________________________
6tisch mailing list
[email protected]<mailto:[email protected]>
https://www.ietf.org/mailman/listinfo/6tisch




--
_______________________________________

Thomas Watteyne, PhD
Research Scientist & Innovator, Inria
Sr Networking Design Eng, Linear Tech
Founder & co-lead, UC Berkeley OpenWSN
Co-chair, IETF 6TiSCH

www.thomaswatteyne.com<http://www.thomaswatteyne.com>
_______________________________________
_______________________________________________
6tisch mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6tisch

Reply via email to