Re: [Lwip] Discussion about IoT Device Classes

[Hannes Tschofenig <[e-mail]>]

Hi Oliver,

your remark illustrates the need to discuss what functionality we
expected to be implemented on these devices.

Without a list of specs the RAM/flash size indication is not terribly
meaningful.

Ciao
Hannes


On 02/03/2017 12:35 PM, Oliver Hahm wrote:
> Hi Hannes!
> 
> On Wed, Feb 01, 2017 at 07:54:03PM +0100, Hannes Tschofenig wrote:
>> On 02/01/2017 07:32 PM, Michael Richardson wrote:
>>>     > PS: I am not sure about the "motes". Are you talking about BLE beacons?
>>>
>>> It's a good example. But really anything that basically can never be field
>>> upgraded.
>>
>> I guess I am fine calling them class 0 devices that are basically
>> outside our scope since they do not run IP.
> 
> I'm not so sure we should discard these devices so hasty. Sure, with just some
> tens of bytes of RAM, you won't be able to implement 6lowpan-HC or DTLS. But
> as soon as we have some kilobytes of RAM (let's say 5kB as on some MSP430
> motes), simple IPv6 connectivity is possible, if its functionality is well
> defined. As far as I know, current IPv6/6lo specifications do not consider
> unidirectional devices, but is there anything wrong in a simple sensor that
> just broadcasts (multicasts) its sensor values to a hardwired IPv6 address
> (e.g., ff02::2)?
> 
> Cheers,
> Oleg
>