On Thu, Mar 15, 2018, 5:14 PM Nash, George <[email protected]> wrote:

> This is my attempt to describe the difference between piid and pi. (UUID
> is defined by RFC-4122)
>
> piid (Protocol Independent ID) - is intended as an identifier that is
> independent of the piid is required to be a UUID.


Typo? Should be "as a device id that is independent of the protocol..."?

Which implies that a protocol-dependent device ID is possible or at least
allowed. If not, then since di is already a uuid, what is the point of a
piid?

Is this sloppy or far-sighted? Why would we want protocol-dependent dis?

  It is intended to detect if multiple instances of the same device has
> been discovered. One use case you have device `A` discovered by two
> different bridges. So your client it told by two different bridges that
> each bridge has a device. The client can look at the `piid` for the device
> from each bridge and it can conclude that both are the same device because
> they have the same `piid` even though the client is getting reports of the
> device twice over two different connections. The `piid` could be used to
> tell if you have discovered the same device over different protocols. For
> example TCP and Bluetooth.
>

Why isn't the di uuid sufficient for that?

Puzzled,

Gregg
_______________________________________________
iotivity-dev mailing list
[email protected]
https://lists.iotivity.org/mailman/listinfo/iotivity-dev

Reply via email to