From: Craig M Prince (cmprince@cs.washington.edu)
Date: Wed Dec 01 2004 - 00:50:03 PST
Reading Review 12-01-2004
-------------------------
Craig Prince
The paper titled "How to Own the Internet in Your Spare Time" analyzes
various recent high-profile internet worms and their propagation
characteristics. The paper then examines the natural evolution of these
worms to determine how quickly such internet worms can actually spread.
The results are very surprising and worrying -- using the methods proposed
in the paper in the worst case complete propogation takes on the order of
seconds! The paper also mentions several methods for discretely
propogating worms and allowing "patching\upgrading" of worms. Finally, the
authors proposed the creation of a committee to combat worms and worm
outbreaks.
This paper was most interesting because it actual does analysis of how
fast worms can spread on the internet. It also does a good job of
describing the "doomsday" scenario and shows how such a scenario is not
beyond the realm of possibility. I thought that the permutation scanning
technique was especially interesting in that it provides a type of
distributed algorithm in order to avoid repeated work on the part of the
infected machines. This is a rather simple modifcation that could have a
large impact on the time it takes for a worm to spread.
The paper did such a good job of convincing me that worms can be
propogated quickly, that their suggested defenses seemed lack-luster and
hopeless. It seems that someone extremely motivated could quickly cause
widespread outages of the internet and tremendous damage, and there is
really nothing we can do about it. All that would be needed is one small
bug...
The paper's argument for a cyber attack CDC did not seem very covincing.
First, there is a question as to whether this should be part of the
private or public sector (companies exist already who focus on worms and
virus threats). Also, most of the proposed defenses still involved a human
in the loop, which they showed is not effective for a quick attack. Is the
best we can hope for to simply mitigate the damage of attacks
after-the-fact?
This archive was generated by hypermail 2.1.6 : Wed Dec 01 2004 - 00:50:03 PST