Re: Defaults: Getting concrete (round 2) from Wan-Teh Chang on 2013-04-23 (public-webcrypto@w3.org from April 2013)

From: Wan-Teh Chang <wtc@google.com>
Date: Tue, 23 Apr 2013 12:50:44 -0700
To: Richard Barnes <rbarnes@bbn.com>
Cc: Ryan Sleevi <sleevi@google.com>, "public-webcrypto@w3.org" <public-webcrypto@w3.org>
Message-ID: <CALTJjxHAHVOAW2rYfgu-x8r_mrYTMZhNkQnD8CYntZaM=E+aqA@mail.gmail.com>

I think there is a small number of parameters that the low-level API
can provide reasonable defaults.

1. The IV for block cipher CBC and OFB modes.

2. The public exponent for RSA key generation.

As for the nonce for block cipher CTR and GCM modes, there is usually
an efficient way to achieve uniqueness if we know more about the
application or protocol. For example, if when encrypting network
packets, we can use the packet sequence numbers. This is why I think
it is not as useful for the UA to generate default random nonces for
CTR and GCM modes, even though it is easy to do so.

Wan-Teh

Received on Tuesday, 23 April 2013 19:51:15 UTC