Re: [Suit] Fwd: Firmware Update Paper

[Szymon Słupik <[e-mail]>]

Hi Emmanuel,

Thanks fo the detailed info.

Some comments of mine in line below

In the experiment you point out (reported in Table 3 and 4 in [1]) the
> transfer of the firmware binary + the manifest (totaling ~35kB) took about
> 11 seconds.
>

[SS] yeah... that 35kB is a bit unrealistic... for production stacks
(including application functionality) we see the FW sizes (uncompressed) to
be in 200kB range. But I take it the times scale linearly, so for 200kB the
transfer time would be about 1 minute. Have you considered the link
saturation during the transfer? In many cases you do not want the transfer
to [significantly] affect the operation of the network, so a strategy to
reduce the transfer duty cycle may be prudent.


> The 802.15.4 radio was used in 2.4GHz band in default mode (we used an
> off-the-shelf SAMR21 board [2] for this experiment).
>

[SS] And that was a single-hop, right? Have you considered multi-hop
transfers over a mesh network?


> The signature verification was performed with the HACL library (ed25519)
> on an ARM Cortex-M0+ and took approx. 7 seconds -- so yes, pretty long
> indeed, in this case.


[SS] Does that scale linearly with image size? Do memory requirements (RAM
usage) increase with the image size?

Best
Simon

On Tue, Dec 3, 2019 at 10:05 PM Emmanuel Baccelli <
[e-mail]> wrote:

> Hi Szymon
>
> below some more information, since you ask ;)
>
> In the experiment you point out (reported in Table 3 and 4 in [1]) the
> transfer of the firmware binary + the manifest (totaling ~35kB) took about
> 11 seconds.
> The 802.15.4 radio was used in 2.4GHz band in default mode (we used an
> off-the-shelf SAMR21 board [2] for this experiment).
> The signature verification was performed with the HACL library (ed25519)
> on an ARM Cortex-M0+ and took approx. 7 seconds -- so yes, pretty long
> indeed, in this case.
>
> For other microcontrollers and/or with other signature schemes or
> alternative implementations/libraries, the speed & memory footprint we
> measured are compared in Tables 8, 7 and 6 (see [1])..
>
> As observed in the paper:
>     - the time spent on signature verification heavily depends on the type
> of microcontroller, on the signature scheme, and on the specific
> implementation;
>     - crypto's footprint in Flash memory can significantly impact the size
> of update binaries which need to be transferred over the network.
>
> Cheers,
>
> Emmanuel
>
> [1] https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8725488
> [2] https://github.com/RIOT-OS/RIOT/wiki/Board:-Samr21-xpro
>
> On Tue, Dec 3, 2019 at 9:53 AM Szymon Słupik <[e-mail]> wrote:
>
>> Thanks Emmanuel,
>>
>> So the PHY data rate was 250kb/s, right?
>>
>> WRT Figure 3 (tIme spent) 38% is signature verification and 60% is
>> transfer. That seems like signature verification is taking a huge amount
>> of time... or was it a different data rate?
>>
>> Best
>>
>> Simon
>>
>>
>> On Tue, Dec 3, 2019 at 5:37 PM Emmanuel Baccelli <
>> [e-mail]> wrote:
>>
>>> Hi Szymon
>>>
>>> in the paper [1] for our experiements we used CoAP as transport, over
>>> UDP, 6LoWPAN and IEEE 802.15.4 low power radio.
>>>
>>> However the open source RIOT implementation which stemmed out also works
>>> out-of-the-box on other link layers too (among others BLE [3], wired [1]
>>> ...).
>>>
>>> Cheers,
>>>
>>> Emmanuel
>>>
>>> [1] https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=8725488
>>> [2] https://github.com/RIOT-OS/RIOT/tree/master/examples/suit_update
>>> [3] https://github.com/RIOT-OS/RIOT/pull/12391
>>>
>>> On Tue, Dec 3, 2019 at 12:58 AM Szymon Słupik <[e-mail]> wrote:
>>>
>>>> Hi Hannes,
>>>>
>>>> What transport did you use? I could not find that information
>>>> explicitly stated in the paper...
>>>>
>>>> Thanks!
>>>>
>>>> Simon
>>>>
>>>> On Tue, Dec 3, 2019 at 12:20 AM Brendan Moran <[e-mail]>
>>>> wrote:
>>>>
>>>>> Hi David,
>>>>>
>>>>> The manifest generator that I released previously was apache
>>>>> 2.0-licensed. I don't plan to change that for the next release. I'm also
>>>>> developing a manifest parser which I anticipate being released under the
>>>>> same license.
>>>>>
>>>>> Best Regards,
>>>>> Brendan
>>>>>
>>>>> -----Original Message-----
>>>>> From: Suit <[e-mail]> On Behalf Of David Brown
>>>>> Sent: 29 November 2019 18:36
>>>>> To: Emmanuel Baccelli <[e-mail]>
>>>>> Cc: [e-mail]
>>>>> Subject: Re: [Suit] Firmware Update Paper
>>>>>
>>>>> On Fri, Nov 29, 2019 at 12:46:42PM +0100, Emmanuel Baccelli wrote:
>>>>>
>>>>> > The open source implementation stemming from our paper [1] is
>>>>> embedded
>>>>> > into the RIOT operating system, which is indeed licensed with LGPL.
>>>>> > The implementation is compliant with draft-ietf-suit-manifest-00 and
>>>>> > has recently been merged into the main branch of RIOT, see [2].
>>>>> > Reuse and further contributions to this code base are welcome!
>>>>> >
>>>>> > Relicensing this code is not planned as far as I know (@ code
>>>>> co-authors:
>>>>> > please correct me if I'm wrong).
>>>>>
>>>>> This is fine.  The authors of the code are free to license the code
>>>>> however they wish.  I just want to make sure it is clear that this code is
>>>>> not useful as a general example, and if the SUIT wishes to have reference
>>>>> code, it will need to be licensed differently.
>>>>>
>>>>> > Related: we know of several companies, big and small, which use RIOT
>>>>> > in their IoT products (and thus use software including -- but not
>>>>> > limited to -- LGPL
>>>>> > code) and they are quite happy with it.
>>>>>
>>>>> I'm sure there are a few companies that are willing to use LGPL
>>>>> licensed embedded code, but their existence doesn't negate that there are
>>>>> large numbers of users who will be unable to use (or even look at) this
>>>>> code.
>>>>>
>>>>> One example is that both Zephyr and MCUboot are licensed under the
>>>>> Apache 2..0 license.  Most parties feel that the Apache 2.0 and the LGPL 2
>>>>> license are incompatible, and this code cannot be linked together into a
>>>>> single product.  Since I'd like to include SUIT support into MCUboot, this
>>>>> means I'll have to be doing an implementation from scratch.
>>>>>
>>>>> It's not my place to argue about the licensing of RIOT, but I would
>>>>> like to see reference code for SUIT that can be used as widley as possible,
>>>>> and the licensing of this particular code prevents it from being used for
>>>>> that.
>>>>>
>>>>> David
>>>>>
>>>>> _______________________________________________
>>>>> Suit mailing list
>>>>> [e-mail]
>>>>> https://www.ietf.org/mailman/listinfo/suit
>>>>> IMPORTANT NOTICE: The contents of this email and any attachments are
>>>>> confidential and may also be privileged. If you are not the intended
>>>>> recipient, please notify the sender immediately and do not disclose the
>>>>> contents to any other person, use it for any purpose, or store or copy the
>>>>> information in any medium. Thank you.
>>>>>
>>>>> _______________________________________________
>>>>> Suit mailing list
>>>>> [e-mail]
>>>>> https://www.ietf.org/mailman/listinfo/suit
>>>>>
>>>> _______________________________________________
> Suit mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/suit
>