Security software performs poorly in exploit test
- — 14 October, 2008 10:46
Security software suites are doing a poor job of detecting when a PC's software is under attack, according to Danish vendor Secunia.
Secunia tested how well a dozen Internet security suites could identify when a software vulnerability was being exploited, said Thomas Kristensen, Secunia's CTO.
That's a different approach from how the programs are architected today. Security software tends to focus on the detection of malicious software that ends up on a PC after a vulnerability has been exploited. The software is updated with signatures, or data files, that recognize certain malicious payloads that are delivered to the PC post exploit.
There's a distinct advantage in focusing on detecting an exploit rather than defending against innumerable payloads, Kristensen said. The exploit itself doesn't change and must be utilized in the same way for the PC to be hacked.
An innumerable number of payloads -- ranging from keystroke loggers to botnet software -- could be deployed during an attack against a vulnerability.
Identifying the exploit isn't easy work, however, Kristensen said. Versions of the program affected by the vulnerability must be analyzed before and after a patch is applied in order to figure out how the exploit works.
For its test, Secunia developed its own working exploits for known software vulnerabilities. Of those exploits, 144 were malicious files, such as multimedia files and office documents. The remaining 156 were exploits incorporated into malicious Web pages that look for browser and ActiveX vulnerabilities, among others.
Symantec came out on top, but even then, its results were not stellar: the company's Internet Security Suite 2009 detected 64 out of 300 exploits, or 21.33 percent of the sample set.
The results then became much worse. BitDefender's Internet Security Suite 2009 build 12.0.10 came in second, detecting 2.33 percent of the sample set. Trend Micro's Internet Security 2008 had the same detection rate as BitDefender, followed by McAfee's Internet Security Suite 2009 in third at 2 percent.
Kristensen cautioned that Secunia was aware most vendors are not focused on detecting exploits. But it would benefit vendors to start creating signatures for exploits rather than merely payloads, since it could save them more time. There are far fewer exploits than payloads, he said.
Vendors such as Symantec appear to be moving in that direction, as it has created signatures for Microsoft-related exploits, Kristensen said.
"We are not seeing any of the other vendors having anything similar to that," Kristensen said.
In the meantime, users should apply software patches as soon as those patches are released. If there's a delay between when an exploit is public and a patch released, users also can simply avoid using the particular program.
"Too many people think they have nothing to be worried about if they only have [antivirus software]," Kristensen said. "Unfortunately, that is definitely not the case."
Anonymous
Wed 15/10/2008 - 13:37
It's protection against in the wild threats that really counts
It's great that Secunia is prepared to take the time to develop potential security exploits and then test the various security products against them.
They do find vulnerabilities, but 99% of them are bits of software that will never see the light day, let alone impact anyone living in the real world. Every now and then, they find a 1%-er that's important. But you have to remember they're testing products against laboratory developed viruses as opposed to in the wild viruses.
For a real-world analogy, it's like saying that it'd be devastating to be hit by a meteor so therefore everyone should go build meteor shelters.
It's much more important to make sure you know what's really impacting people, and then protect against that.
Best Regards, Lloyd Borrett
Marketing Manager, AVG (AU/NZ)
www.avg.com.au