Re: DoS attack ?

[John C Klensin <klensin@jck.com>]

Exactly.  And, as with the whois case, it is reasonable that
database providers be able to protect themselves.  The question
about the reasonable limits of such protection is a complicated
one (going back to the dawn of shared/ commercial databases and
information retrieval and query systems), but, fortunately, is
rarely or ever a protocol problem.

    john

--On Thursday, 06 December, 2001 12:51 -0500 Eric
Brunner-Williams in Portland Maine <brunner@nic-naa.net> wrote:

> YangWoo,
> 
> I wouldn't refer to this as a DoS attack, rather as a
> mechanism for the (unauthorized) replication (of some or all)
> of the store.
> 
> It may matter that what amounts to an unauthorized zone
> transfer ties up the data flow source node, or its local
> bandwidth, and that these may be synchronized across multiple
> sink nodes to fully consume either host cpu or subnet i/o
> resource, but that is only one case (discard-at-sinks) of
> resource replication.
> 
> We see the latter in data mining against whois servers, and I
> would expect that resource capture isn't confined to
> telemarketers mining for telephone contact data.
> 
> Eric
> 
> 
>