Re: [Suit] self-describing format vs fixed/binary manifest structure

[David Brown <[e-mail]>]

On Thu, Dec 13, 2018 at 11:22:07PM +0000, Martin Pagel wrote:

>In general I actually agree with your assessment. If you have many
>options, a well defined format like CBOR is much better than custom C
>structures as you can use a standard, already vetted parsing library.
>The standard parser library turns the CBOR data into a data
>structure, in IoT usually a pretty complex C structure. As the
>generic parser does not understand the semantics of the fields (only
>the syntax as defined by the CDDL), the developer still has to write
>code to verify a complex data structure and enforce semantics and you
>still have to test/verify such code.

There has been some good discussion on list about pull vs push
parsing.  What you are describing is known as a push parser.  The
parser parses the CBOR and converts it into a fairly rich data
structure that describes the contents of the CBOR.  This is a common
approach, and is quite useful on "powerful" computers, such as
host-side tooling to process the data.

However, in a highly-resource-constrained device, this is not
typically how it is done.  Usually a "pull" parser is used, which will
have entry points to effectively expect a specific (and likely
limited) variant of the manifest, and simply abort if it see what is
unexpected.  The data is not converted to another structure, but
directly parsed/used in place.  For this kind of use, CBOR works very
well.  Draft 3 of the manifest document has an accompanied example
parser that works this way.

As a comparitive example, MCUboot uses a custom encoding format for
the data in its manifest.  It consists of a small amount of data in a
fixed structure (with defined byte ordering and packing), and a set of
tag-length-value items (effectively a single CBOR map, with integer
keys and all values as bstrs).  This is parsed with a similar custom
parser for the data expected.  I expect the parser for the SUIT
manifest to be somewhat larger than this, but within a similar
magnitude.  Even this simple case of a single map, there is enough
variation that a plain struct would not be appropriate to represent
this information.

David