- In 2005 USB drives became popular and malware started using them to propagate.
- Even three years after malware started actively using this method to infect customers, Microsoft refused to accept the reality of the problem and continued offering AutoRun enabled by default in the Windows OS’s. However in 2008 Microsoft added an option for disabling AutoRun via policies or manual registry entries. However the workaround provided did not work. Even when disabled users were still open to attack from the AutoRun infection vector.
- In July 2008 Microsoft published MS08-038 which “fixed the broken fix” but this was only available via Windows Update for Windows Vista and Windows 2008. Instead of patching XP users as well, it kept the problem unsolved in what some might consider a business strategy to sell more Vista licenses.
- Towards the end of 2008 Conficker showed up taking advantage of the AutoRun feature in a never seen before manner. It created an autorun.inf file whose content looked like garbage yet was fully functional. All the Microsoft recommended workarounds to date via NoDriveTypeAutorun policies continued to be useless against malware exploits.
- In early 2009 and due to Conficker’s success Microsoft corrected a bug (CVE-2009-0243) which fixed portions of the previous problem and which was pushed out automatically to all Windows XP users. Amazingly it wasn’t considered a “security patch” and does not have an associated Microsoft Bulletin. In addition the patch modified the behaviour of AutoRun and after applying it created a new registry entry which was required to be manually configured correctly. Effectively AutoRun continued being a problem for the vast majority of users.
- In mid 2009 there seems to be some light at the end of the tunnel and Microsoft decides to improve the security of AutoRun in writeable removable media by preventing the AutoPlay dialog window in USB drives. However this is only included by default under Windows 7. Windows XP users, still the most widely used platform by far, had to manually download and install KB971029. This move was effectively useless from the point of protecting XP users from malware infection. Again some might consider this move a business-driven decision to “keep security low in XP in order to drive sales of the more secure Windows 7″.
- In July 2010 Stuxnet shocks the world. It is able to propagate via USB drives without requiring an autorun.inf file and using a zero-day vulnerability in .LNK files which allows for code execution even with AutoRun and AutoPlay disabled, which Microsoft promptly patches.
- Finally in February 2011 Microsoft decided to push an update to fix the problem for Operating Systems prior to Windows 7.
Microsoft recently started installing its Microsoft Security Essentials (MSE) free antivirus product via the Operating System update mechanism to computers which don’t already have an antivirus installed. Basically Microsoft is saying they are worried about the security of its users and they need to make sure they are protected. Perhaps Microsoft is trying to position itself as a provider of secure Operating Systems given the market perception of Linux, Apple and potentially Google as having more secure alternatives to Windows OS, but that’s a different story.
We agree with Microsoft; it’s better to have some protection than not having any at all. However the way the guys in Redmond are executing the idea is risky from a security perspective and could very well make the malware situation much worse for Internet users. That’s why we encourage Microsoft to continue using Windows/Microsoft Update but instead to push all free antivirus products available on the market, not just MSE.
These are the reasons why pushing only MSE from Windows/Microsoft Update is a very bad idea:
- MSE is not a good solution to the malware problem. While the argument of protecting users who do not have AV is commendable, the reality is that MSE only installs on computers with a valid Windows OS license (paid to Microsoft).
- The problem is that an estimated 40% of worldwide computers connected to the Internet are running pirated software and spreading viruses, especially in China, Latin America, Asia, Southern Europe, etc. So while Microsoft wants us to think it is doing this out of the goodness of their hearts, the reality is that the measure will have little impact as millions and millions of unlicensed Windows PCs will continue spreading viruses and infecting the rest of us.
- Even Microsoft itself acknowledges that malware infections are more prevalent in illegal copies of Windows: “There is a direct correlation between piracy and the malware infection rate” said Jeff Williams, the principal group program manager for the Microsoft Malware Protection Center. If that’s correct and the objective is truly to protect users from malware, then why doesn’t Microsoft allow MSE to install in pirated copies of Windows OS?
- Monocultures are a hacker’s paradise. If pushing MSE via Windows/Microsoft Update is very successful it will end up creating a monoculture of hundreds of millions of users having the same antivirus product. Right now hackers have to worry about bypassing multiple antivirus products and protection layers every time they release a new piece of malware. Having to bypass only one AV product makes their life so much easier. This alone will allow hackers to push more new malware that bypasses MSE exclusively and infect many more users with every new variant. Alternatively, reverse engineering of MSE and related Windows components will boom, potentially discovering zero-day vulnerabilities which could cause infections in tens of millions of PCs with a single attack. Monoculture in Operating Systems is in and by itself bad. Monoculture in security is A VERY BAD THING.
- Insufficient Detection. Even though MSE is a good basic product, from a detection perspective it has not proven itself to provide sufficient protection according to the latest independent comparative studies:
- AV-Comparatives.org’s latest On-Demand Test ranks MSE 15 out of 20 in signature detection while vendors with alternative free antivirus products were ranked well above that.
- In AV-Test.org’s latest Real-World Test MSE could not achieve the minimum score to obtain certification, while vendors with alternative free antivirus products did. MSE was ranked as one of the worst three products.
- Not Enough Prevention. There are other free antivirus alternatives on the market which offer much more than just reactive signature detection. These more advanced (and still completely free) products have multiple security layers which provide users with proactive protection, such as web filtering, behavior blocking, instant messaging filters, etc. MSE provides very basic antivirus protection, certainly not enough to protect users against today’s malware threat landscape.
- Secure the Operating System itself. Even though Microsoft has made significant improvements in securing the OS in recent years, there is still a long way to go as witnessed by the constant zero-day vulnerabilities that are published every month, such as the incredibly dangerous LNK vulnerability that Stuxnet exploited. Microsoft’s security resources should work on making the OS more secure, not just putting a band-aid on it. Who knows, maybe someday if Microsoft manages to really make their OS secure, antivirus products won’t be needed anymore. But until that day comes, Microsoft should make a serious development effort to secure the OS from the ground up and not limit the security tools currently available to its users.
In summary, while it’s commendable that Microsoft is trying to protect users, offering only “their” basic MSE antivirus provides neither sufficient protection against today’s threats nor does it solve the malware problem of millions upon millions of pirated PCs who will continue spreading viruses. In fact, it can easily achieve the contrary by making it easier for hackers to infect users. Microsoft should offer the complete portfolio of more advanced and secure alternatives of free antivirus products and time-limited versions of paid security suites, allowing users to choose any of them from the Optional Windows/Microsoft Update.
Finally the full report of the comprehensive Full-Product Test from German independent antivirus tester AV-Test.org is out.
Panda Internet Security has received excellent scores in all categories, accomplishing top rank along with two other vendors. According to Andreas Marx, CEO of AV-Test.org, “Panda Internet Security was one of only three products which was able to receive the highest scores during this exhaustive test which was performed over a period of 12 weeks“.
The Full-Product Test is a very extensive test which looks at many different aspects of a security solution:
- Real-World Testing – protection against 0-day and web/email malware
- Dynamic (Behaviour) Detection Testing – blocking of malware on execution
- Detection of Large Malware Collection – testbed from last 3 months’ malware
- Detection of Widespread Malware – based on WildList criteria
- Repair and removal of widespread malware
- Removal of malicious components and remediation of system modifications
- Detection of hidden active rootkits
- Removal of hidden active rootkits
- Average slow-down of the computer
- False positives during static on-demand scanning
- False positives during dynamic on-access scanning
Some additional comments from AV-Test.org regarding Panda Internet Security:
Panda Internet Security showed impressive high results for the static and dynamic detection of new malware.
The detection and removal of actively running stealth malware such as rootkits was no problem for Panda Internet Security, but for many other reviewed products.
We tested not only the protection against known and unknown malware, but also the removal of critters which had previously infected the system and Panda Internet Security received 5.5 out of 6.0 possible points in these two category, the highest scores archived by a program during this exhaustive review.
Not only the protection against and removal of new malware was very high, but at the same time Panda Internet Security had less impact on the system from the usability point of view.
Here is yet another example of a company distributing malware to its userbase. Unfortunately it probably won’t be the last.
Today one of our colleagues received a brand new Vodafone HTC Magic with Google’s Android OS. “Neat” she said. Vodafone distributes this phone to its userbase in some European countries and it seems affordable as you can get it for 0â‚¬ or 1â‚¬ under certain conditions.
The interesting thing is that when she plugged the phone to her PC via USB her Panda Cloud Antivirus went off, detecting both an autorun.inf and autorun.exe as malicious. A quick look into the phone quickly revealed it was infected and spreading the infection to any and all PCs that the phone would be plugged into.
A quick analysis of the malware reveals that it is in fact a Mariposa bot client. This one, unlike the one announced last week which was run by spanish hacker group “DDP Team”, is run by some guy named “tnls” as the botnet-control mechanism shows:
00129953Â |.Â 81F2 736C6E74Â Â Â Â Â Â Â Â |XOR EDX,746E6C73 ; â€tnlsâ€
The Command & Control servers which it connects to via UDP to receive instructions are:
Once infected you can see the malware “phoning home” to receive further instructions, probably to steal all of the user’s credentials and send them to the malware writer.
Interestingly enough, the Mariposa bot is not the only malware I found on the Vodafone HTC Magic phone. There’s also a Confiker and a Lineage password stealing malware. I wonder who’s doing QA at Vodafone and HTC these days
As you will notice we’ve migrated the Panda Research blog to a new platform, which I’m hoping will take less time to manage, specially moderating comments and filtering spam, which took a lot of time with the cumbersome Microsoft blogging platform.
If you’ve posted a comment which hasn’t made the migration, please post it again. I’ll try my best to moderate it as soon as possible.