Fighting malware and spam

Vulnerability test raises hackles

Secunia suite trial slates lack of PoC detection, but test methods called into question.

Vulnerability specialist Secunia published the results of a trial of internet security suites this week, with stark findings showing very poor detection rates of the selection of exploits and proof-of-concept overruns used. However, the test has run into considerable criticism from several angles.

Secunia's test focused on 300 known vulnerabilities, around half of them based on file types and the other half based on web pages. These were represented mostly by payload-free proof-of-concept attacks designed to trigger vulnerabilities, and in some cases prove that code execution could be achieved, while some fully fledged exploits were also included. A dozen 'Internet Security' suites were lined up and set to scan the files and web pages in question, and of them only Symantec's Norton product put in any kind of a performance, achieving just over 20% of all samples and just over 30% of those deemed 'important'. No other product detected more than 2.5% of the full testset, or better than 4% of the most dangerous items.

In the wake of the test's publication, a range of criticisms have been levelled at Secunia, mostly focusing on the test methodology, which many claim failed to properly exercise any of the protection features included in these suites beyond the basic anti-virus scanner. Many product developers have claimed that if the vulnerabilities had actually been triggered, rather than scanned statically, other technologies would have kicked in to block any resulting exploits, and that in many cases payloads would be more likely to be blocked than the first stages of exploitation.

Others have criticised the lack of clarity and openness about how the testing was carried out, and some have seen the whole business as a stunt to promote Secunia's own vulnerability scanning products and services. A blogger at Kaspersky Lab points out that it is a far better policy to patch vulnerabilities as soon as they are discovered, rather than leaving holes open and relying on other software to block anything that tries to wriggle through, an angle which Kaspersky's integrated vulnerability scanner, like those offered by eEye Blink and BitDefender, attempts to cover. Secunia representatives themselves claim, in a blog entry announcing their results, that there is a serious lack of proper and timely patching among home users and businesses alike.

Full details of the Secunia test are published in a PDF here, with some comment from the company here. Further comment is in The Register here, on the Sunbelt blog here or at Panda here.

17 October 2008

Tags: exploit, secunia, test, vulnerability.   

1 comment

The Kaspersky guy makes a pretty good point - if you have a broken window in your house, you don't hire someone to sit on guard with photos of all known burglars, you just fix the window. Patching known vulns makes a lot more sense than expecting AV to have sigs for all possible attempts to exploit them (although I'd expect them to do pretty well with the payloads).

Secunia seem to have taken a bit of a pasting with this, they could probably have found a better way of promoting the idea of vuln management and patching - but a good scandal always helps with brand recognition...

by Paul Westlake, 20 October 2008, 09:34

Comments are closed.