Re: [lisp] Call for interest in working on security

[Sam Hartman <hartmans-ietf at mit.edu>]

I really appreciate your thoughtful comments.

I'll admit that I am personally interested in the non-alt parts of
LISP security.  My perhaps naive assumption is that something like
SIDR can be applied and that I'd like to get a better handle on the
surrounding landscape before I think in detail about that.

However thinking about how the alt differed from traditional
BGP--particularly in terms of the implications of high aggrigation on
security--is something I had not done.

One other comment:
>>>>> "Eliot" == Eliot Lear <lear at cisco.com> writes:
    Eliot> I will make the following two general assertions, as relates to your comment
    Eliot> about key management:

    Eliot>   • the problem is operationally difficult when there are large numbers of
    Eliot>     signers and/or long trust chains.
    Eliot>   • a small number of signers requires concentration of trust far beyond the
    Eliot>     point of what exists today (perhaps by 3 orders of magnitude or more),
    Eliot>     leading to potential continuity problems.

    Eliot> These were the trade-offs I explored with NERD.


These are interesting.  However I don't think I have talked about key
management for anything involving signers at all so far.  I've talked
about key management for map register, but there explicitly talked
about a solution with no public key ops.