Security patch approach is failing
If MS Web admins can't keep up to date
A noted security expert has said current security practices are too reliant on expecting users to apply patches and has suggested better monitoring might lead to more robust security.
Bruce Schneier, chief technology officer of Counterpane Internet Security, said the outbreak of the Code Red Worm, which targets vulnerable IIS Web servers, shows that "the patch treadmill doesn't work".
Schneier argues that when even Microsoft's Web admins can't keep up to date with patches (one of the many sites defaced by the worm was the Windows Update page) it shows the approach is failing. He said the patching approach doesn't take into account human weaknesses or that patches sometimes break other parts of a network or sometimes require for critical systems to be taken offline in order to be applied.
These are good points, but can the defenders of networks come up with a better approach (and preferably one that doesn't blame the victim for security breaches)?
Schneier certainly thinks so and advocates wider use of security monitoring as a means to fill the security gaps.
"If you are monitoring your network carefully enough, you'll catch a hacker regardless of what vulnerability he exploited to gain access," said Schneier.
"Monitoring makes a network less dependent on keeping patches up to date; it's a process that provides security even in the face of ever-present vulnerabilities, uninstalled patches, and imperfect products."
Schneier admitted that vigilant monitoring does not "solve" computer security, but his argument that is a much realistic way of providing resilient security is worth considering.
The reactive nature of monitoring can give attackers time to do some serious damage, so we can't see the approach will take us away from the need to apply security patches altogether, but it has the potential to reduce risk.
Firewalls alone don't provide adequate defences, particularly against something like Code Red, which is a pre-programmed worm that unleashes a distributed attack against a predetermined target, and intrusion detection systems are generally only as good as their latest attack signatures.
We can't see that monitoring would be much good in isolation but it might well be successful at picking up problems more effectively and making Internet security less fragile. After all, it can hardly hurt... ®
Internet survives Code Red
Code Red bug hits Microsoft security update site
IIS worm made to packet Whitehouse.gov
IIS buffer-overrun attack has been scripted
MS confronts another IIS system-level hole
Yet another IIS exploit reported
Microsoft IIS hole gives System-level access
Security patch distribution - it's Trojan time