In message <[EMAIL PROTECTED]>, Robert Stone writes:
>On Thu, Mar 22, 2001 at 10:17:21AM -0600, Steven M. Bellovin wrote:
>> That said, an RFC that discussed DoS avoidance strategies would be
>> a good idea. I'm agnostic about whether that should be done in a
>> WG or as an individual submission, but BCP status would be a good
>> one to aim for. How this process should be organized is up to the
>> AD's and the IESG. (Also note that the next rev of draft-rescorla-sec-cons
>> has a good section on DoS attacks.)
>
>For what it's worth, the following document could provide a starting point:
>
>http://www.cert.org/reports/dsit_workshop-final.html
>
>It provides some specific recommendations, but is too vague in some areas
>and is, perhaps, a little out-of-date.
>
That report is specific to bandwidth-consumption attacks. I was suggesting
a broader document for protocol designers that discussed CPU attacks,
memory attacks, spoofing, black hole creation, etc., as well as design
principles for minimizing or preventing such problems. For example, by
not creating state until the third message of a four-way handshake, you
minimize the chance of memory consumption attacks via spoofed IP
addresses. (SCTP does that.) On the other hand, although cryptography
is a powerful tool against intrusions, it can be expensive; an enemy
might try to force you do do more cryptographic calculations.
--Steve Bellovin, http://www.research.att.com/~smb
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
