Re: root zone signing and key lengths/lifetimes

[Bill Manning <bmanning@isi.edu>]

% "Scott Rose" <scottr@antd.nist.gov> writes:
% 
% Hi Scott,
% 
% I'm trying to catch up, sorry for being so late.
% 
% > Mainly I was wondering why the draft set up frequency of key
% > rollovers.  Not that it's a huge technical problem.  Although the
% > human nature side of me is reminded that the more frequent the
% > operation, the more frequent human error creeps in.
% 
% Furthermore I'd much prefer to force a certain beat to the rollovers
% to get experience with the mechanism and whether it works or not.  
% 
% To have the first ever rollover of the root key 12 or 18 months into a
% possible large scale DNSSEC deployment (if there will ever be such a
% thing) is a rather scary concept to me.
% 
% Johan


Not that this is germaine to most of the folks on this list but there is
a little testbed that is investigating key mgmt in the context of DNSSEC/DS.
The testbed has rolled the root key three times in six weeks (mostly due to
things like making the mistakes of using the mandatory to implement (DSA), then
the fact the default RSA is MD5.  We are now at RSA/SHA1 with an "fat" key
size.  The expectation is to roll a new key on about that schedule.

Key rollover is problematic, esp from the point of view of the endsystem
since the current technology requires replacing all the distributed copies
of the key. :(   (see previous Bush/Andrews thread on possible ways around
this problem)

--bill