Seven IE 9 Security Recommendations for Microsoft

1
2
3
4
5
6
7
8
1 of 9

Seven IE 9 Security Recommendations for Microsoft

by Brian Prince

2 of 9

Sandboxing Technology

Microsoft could improve things in IE 9 by adopting some of the sandboxing approaches Google uses in its Chrome browser. IE 9 has "Protected Mode," which is similar, though not designed for the same purpose, said Aaron Portnoy, TippingPoint security research team lead.

3 of 9

Plug-ins Out of Process

"I believe it would be beneficial to IE's security posture for it to run as many third-party plug-ins out of process as possible," TippingPoint's Portnoy said. "By running them in-process, an attacker can utilize known or unknown techniques to defeat or weaken exploit mitigations such as DEP [data execution prevention] and ASLR [address space layout randomization]."

4 of 9

Memory Randomization

By randomizing memory addresses used by popular functions, attackers will have a tougher time identifying and repeating exploits against vulnerable code, said Rick Moy, president of NSS Labs.

5 of 9

Redirect Hopping

"Drive-by downloads make use of multiple redirects to confuse reputation systems [such as IE SmartScreen and Google SafeBrowsing] and bring the user to an unwanted page with an exploit," NSS Labs' Moy said. "Disallowing more than one sequential redirect could significantly increase the effectiveness of reputation systems."

6 of 9

Content Security Policy

By implementing content security policy, Microsoft can offer users additional protections against cross-site scripting and click-jacking. Mozilla has already begun work in this direction for its Firefox browser.

7 of 9

Plug-in Registry

Moy said he would like to see users get help differentiating between good and bad plug-ins. "A combination of code hashing/white listing and reputation could help potential users know who made and packaged the application, and what their track record is," he said.

8 of 9

Secure API for Plug-ins

"Browsers should take the lead in protecting plug-ins from memory-based attacks, such as buffer overflows and heap sprays," Moy said. "Providing a secure API instead of direct memory access would go a long way toward reducing the attack surface."

9 of 9

No Title

Top White Papers and Webcasts