Anti-virus tests find security programs fooled by attack vector

Malware that is blocked one time may be allowed through via another route, such as through a local network fileshare or USB drive

A new round of anti-virus testing has found some products fail to detect malware that tries to infect a computer via a different attack vector, such as through a local network fileshare or a USB drive.

The tests, conducted by NSS Labs, sought to find out how effective security products are at detecting malware from various attack vectors. Malware can be delivered to a computer via rigged websites, email attachments, and USB flash drives, among other ways.

[ Learn how to greatly reduce the threat of malicious attacks with InfoWorld's Insider Threat Deep Dive PDF special report. | Stay up to date on the latest security developments with InfoWorld's Security Central newsletter. ]

Although drive-by downloads remain the most common attack vector, about 15 percent of attacks are delivered via email with a malicious attachments, such as a PDF document.

Many security products allow users to download all of their email to their inbox by default and not scan it, even if it contains malware.

"Surprisingly, many products tested did not remove malware from the inbox by default," according to the report, titled "Socially-engineered Malware Via Multiple Attack Vectors."

Of the 10 products tested, the average protection rate was just 36 percent. NSS Labs said that if a company runs a centralized, server-based security product that is integrated with the email servers, such as Microsoft's Exchange or IBM's Lotus Notes, the malware may be removed before it reaches an end user.

But NSS Labs did find that those products that did not scan email before it arrived in an inbox would scan it if the user decided to save the attachment. That improved the average protection rate, which measured 74 percent, NSS Labs said.

Another possible infection vector is file servers, commonly used in organizations to allow access to documents among users. But those files servers can become repositories for malware, allowing bad programs to proliferate among a high number of users.

"While file servers should have their own anti-malware scanning, this often is not the case, and users must rely on local anti-malware security products to detect the downloaded files," the report said.

About 70 percent of the malware was caught by the 10 products when downloaded from a file server, NSS Labs found.

The strongest aspect of most endpoint anti-virus products is their ability to block malware as it is executed and quarantine it. NSS Labs found that even if malware did make it on a PC, most products performed well at containing it.

"Every vendor product, with the exception of Panda, blocked more malware during execution than by analyzing the entry vectors," according to NSS Labs. "Trend Micro, McAfee, and Sophos lead the group."

But one attack vendor where most security companies are still lacking is detecting malicious payloads that are written only to memory, also known as single-use malware. Malware can, for example, masquerade as a permitted DLL (Dynamic Link Library), which skirts around DEP (Data Execution Prevention) security features in OSes.

"This type of attack circumvents protections that lack behavioral analysis for these attacks," NSS Labs wrote. Only three products from Kaspersky, McAfee and Sophos have features to protect against that style of attack.

NSS Labs, which does not accept money from vendors for its testing, is selling the report on its website for $995.

Send news tips and comments to jeremy_kirk@idg.com.

Recommended
Join the discussion
Be the first to comment on this article. Our Commenting Policies