Re: [Teep] JSON/JOSE vs. CBOR/COSE

[Michael Richardson <[e-mail]>]

Brendan Moran <[e-mail]> wrote:
    > I also support using only one encoding, and I second keeping the TCB as
    > small as possible, so I support using CBOR/COSE as the only
    > encoding. In particular, if TEEP is going to keep a dependency on SUIT,
    > this means that using JSON/JOSE for TEEP requires supporting both
    > JSON/JOSE and CBOR/COSE, since SUIT already has a single encoding:
    > CBOR/COSE. As such, using JSON/JSOE for TEEP dramatically increases the
    > TCB from a single parser & cryptographic wrapper to two of each.

This is true from a generality of specification point of view, but are there any
ways to use TEEP without SUIT?  (I think so, but I could be wrong, as my
TEEP-fu is weak)   It could be that those uses could specify JSON/JOSE, while
all SUIT/TEEP uses could specify CBOR/COSE.

Having said this, I'm already in favour of CBOR/COSE only, I just want to
make sure that we are clear about the possibilities.  A client implementation
could support one or the other, but not both.

    > Personally, I don’t find CBOR/COSE significantly harder to debug than
    > JSON. I find JSON unreadable unless it’s been formatted. I find CBOR
    > unreadable unless it’s been decoded into diagnostic notation. I don’t
    > find a substantial difference in overhead between these two.

+1

--
Michael Richardson <[e-mail]>, Sandelman Software Works
 -= IPv6 IoT consulting =-