[Cfrg] ASN.1 in IEEE 1609.2 (was Re: Curve selection revisited)

[William Whyte <[e-mail]>]

Hi all,

> [Hannes Tschofenig]
>> PS: Regarding the ASN.1 format I am wondering whether the IEEE has done
>> some more detailed analysis of the over-the-wire and the software
>> implementation cost. This would be useful for us in the ACE working group.
>
> [Bob Moskowitz]
> Actually, I mispoke a bit on this.  ASN.1 can be a syntax to describe data.
> It is the BER that defines what is really in the object.  So 1609 as
> developing a task to create an ASN.1 discription of their format and the
> specific BER to create it.  ETSI will be much happier once this is done.

I'm the editor of 1609.2, which currently describes the data
structures in a TLS-like notation and is considering moving to ASN.1.
We've analysed both relative packet size and relative encoding speeds
with ASN.1 (with various different encoding rules) and with the
current encoding. The results are still a bit rough so I can't
distribute them publicly, but we anticipate that they'll be
publishable in a month or so. (I don't know if they're relevant to
this list -- let me know where else I should send them if this list
isn't the place). The high-level outcome, though, is that if you pick
a good set of encoding rules (Octet Encoding Rules, or OER, rather
than the BER/DER that's used in X.509 certs and bloats nested
structures by putting tag and length at every level), then encoding
speed and encoded size are competitive with the current 1609.2 custom
format -- slightly larger and very competitive in speed -- and we
understand ASN.1 has advantages in tool availability, maintainability,
extensibility, and completeness, as well as being preferred for ETSI
standards and for standards sponsored by USDOT.

I would put Bob's statements slighltly differently:
> ASN.1 can be a syntax to describe data.
ASN.1 is a language used to define data structures.
> It is the BER that defines what is really in the object.
The encoding rules (which there are many of, including BER but also
other more compact ones) describe the encoding and decoding for
transmission.
> So 1609 as
> developing a task to create an ASN.1 discription of their format and the
> specific BER to create it.
Yes -- we're creating an ASN.1 description of our structures, and
we're looking at which set of encoding rules it makes sense to specify
for use.

Hope this helps.

Cheers,

William