nanog mailing list archives

Re: botnets: web servers, end-systems and Vint Cerf


From: Gadi Evron <ge () linuxbox org>
Date: Sat, 17 Feb 2007 19:22:24 -0600 (CST)


On Sat, 17 Feb 2007, Sean Donelan wrote:

On Sat, 17 Feb 2007, Gadi Evron wrote:
Public ISPs have been testing these types of systems for over 5 years.
What sorts of differences can you think of that would explain why public
ISPs have found them not very effective?

Public ISPs have been using walled gardens for a long time for user
registration and collecting credit card information.  So they know how to
implement walled gardens.  But what happens when public ISPs use it for
infected machines?


Many already do, successfully.

When I say many I actually mean I know of 6. 3 of them huge, 3 of them
relatively small.

Interesting use of the word "many."  Many people use Multics.

:))

I know of "many" more that have tested it and returned it to various 
vendors.  There are several tough problems people are still trying
to solve.

Yes, but that is because the successful ISPs currently often implement
their own if they have the resources and R&D power. The really big ones
have it automated, the small ones have it limited to be "activated by an
abuse desk person".

        Gadi.


Current thread: