Re: [Fud] Charter Text

[Hannes Tschofenig <[e-mail]>]

Hi Olaf,

>> - we have the case of software updates, not only firmware updates, so
>> I'd rather we talk about the more general case of software updates. Is
>> there a strong reason against this?
> 
> This is more a question of whether a module is regarded as part of the
> firmware (cf. "firmware package" in the text). Some application-specific
> JavaScript code may or may not fall under this category. (You could,
> e.g., discuss if Ethereum contracts are "software" in this sense. But I
> am not sure that this would be addressed in this WG?)

I believe Emmanuel needs to tell us a bit more about the use case he has
in mind.

The use cases I would like to cover are:

- a developer creates a single firmware image, which may consist of
source code from various parties and may also contain binaries from
third parties.

- a developer creates multiple firmware images (since the device has
multiple microcontrollers that need to be updated).

In this model the update service on a microcontroller gets code from a
single source only (at least for the device it appears so since it does
not know whether an included library was written by a different developer).

What I specifically want to exclude is a model like JavaScript where you
can get code from multiple different sources dynamically during code
execution. Not only does this lead to challenges for code execution on
these constrained devices but it also raises all sorts of security
issues, which are difficult to solve on these low end devices. While
there are isolation techniques available (like the MPU on Cortex M class
devices and TrustZone for v8-M in upcoming devices) they have their limits.

Ciao
Hannes

PS: I consider Smart Contracts also outside the scope of this group.
Maybe an IRTF research group should work on them.

> 
> Grüße
> Olaf
>