The goal of this document is help new XP users survive long enough to do their updates. Many of them cant/wont put up acls/nat/firewalls ... but if they follow the steps listed they have a better chance of successfully downloading and updating their new machine then they will have with OUT these steps. It is not meant as a complete XP hardening document. There are lots of documents that discuss in detail how to harden windows (xp,nt,2k...).
If the person doesn't continue to do acls/nat/firewalls, they'll just get infected after the next hole is discovered. And yes, there are plenty of holes that a firewall/nat box won't fix. Still, better than the user only doing Windows Update on the day of install and never having a firewall... Rob Nelson ronelson@vt.edu
On 5-mei-04, at 0:26, Rob Nelson wrote:
If the person doesn't continue to do acls/nat/firewalls, they'll just get infected after the next hole is discovered. And yes, there are plenty of holes that a firewall/nat box won't fix. Still, better than the user only doing Windows Update on the day of install and never having a firewall...
I object to the idea that requiring a software firewall inside a host is a reasonable thing to do. Why on earth would I want to run an insecure service and then have a filter to keep it from being used? Either I really want to run the service, and then the firewall gets in the way, or I don't need the service to be reachable, so I shouldn't run it. System services should only be available over the loopback address. Now obviously this is way too simple for some OS builders, but we shouldn't accept their ugly hacks as best current practice.
On Thu, 06 May 2004 11:45:23 +0200, Iljitsch van Beijnum said:
I object to the idea that requiring a software firewall inside a host is a reasonable thing to do. Why on earth would I want to run an insecure service and then have a filter to keep it from being used?
You object to it, I object to it... but the fact remains that 95% of the user-accessible CPUs (not counting the embedded market) are running software that you have to do unreasonable things in order to make it anywhere near safe to use....
Either I really want to run the service, and then the firewall gets in the way, or I don't need the service to be reachable, so I shouldn't run it. System services should only be available over the loopback address. Now obviously this is way too simple for some OS builders, but we shouldn't accept their ugly hacks as best current practice.
"Best Current Practice" is *so* divergent from "Currently Deployed Practice" that there's little or no common ground.
participants (3)
-
Iljitsch van Beijnum -
Rob Nelson -
Valdis.Kletnieks＠vt.edu