4 ms·

Is there an article anywhere that clearly explains the actual attack vector? The bash script ultimately used by the malware is a wiper designed to work with an

by kyllo 14y ago

Is there an article anywhere that clearly explains the actual attack vector? The bash script ultimately used by the malware is a wiper designed to work with any Linux distribution, with specific commands for SunOS, AIX, HP-UX distributions, Symantec said. As such, it wipes out the /kernel, /usr, /etc, and /home directories. That wouldn't take a particularly sophisticated bash script. But as far as how the payload was actually delivered to the infected machines, there doesn't seem to be a consensus. This article portrays it as a RAT/botnet style attack via "a central computer providing antivirus protection," which suggests that they may have gained privileged access to an antivirus company's update server and pushed the code out to clients via the antivirus software itself. http://www.guardian.co.uk/technology/2013/mar/22/south-korea-cyber-attack http://www.guardian.co.uk/technology/2013/mar/22/south-korea... Meanwhile, this article attributes it to an executable e-mail attachment with a misleading filename, basically a phishing attack: http://www.infosecurity-magazine.com/view/31475/south-korea-cyber-attacks-likely-sparked-by-phishing-emails/ http://www.infosecurity-magazine.com/view/31475/south-korea-... Some articles described the South Korean government as concluding that the attack came from China, and then later articles say that they backtracked on this and decided it actually came from North Korea. I think this part is heavily politicized, and it's entirely possible that it did come from China but the South Korean government found it more politically expedient to blame it on North Korea instead.

drzaiusapelord 14y ago
Reason #452 why I never run AV on servers unless its a file server. AV is just a vector for either fat fingered definitions that destroy your system or a venue for malware to enter your system. I don't trust Sophos and Kaspersky that much. On the desktop, its a tolerated and perhaps necessary evil, but on servers - no way.
kijin 14y ago
My current understanding is that the malware was delivered as part of an update to some commonly used ActiveX controls. The initial infiltration took place 8 months ago, but the malware was designed to go live at a specific time [1]. Over 70 different pieces of malware were used to infiltrate a wide range of targets. The bash script was apparently used on non-Windows targets, of which there were only a few. EDIT: The single most effective attack vector seems to have been the update server for an ActiveX control called XecureWeb, which is used by several banks and their 20 million customers [4]. Just checked and I have XecureWeb too, in a VM that I only use for online banking. Good thing I have that VM turned off 99% of the time. According to SK news media, the primary reason the SK government suspects NK of the latest attack is because the attackers used some of the same malware, same proxy servers, and same techniques that were used in previous attacks, which the SK government also attributes to NK [2]. But AFAIK there is no indisputable evidence that links those previous attacks to NK, either. So there's a bit of circular reasoning involved in the accusation. The initial accusation of a Chinese attack was based on a Chinese IP address (101.106.25.105) that was used by the attackers. This accusation was rescinded when it was found that one of the banks that were hacked have been using the same IP range in its internal network </facepalm> [3]. But even if this wasn't the case, the case for a Chinese attack would be weak because anyone can use a Chinese open proxy. There was also mention of an North Korean IP address (175.45.178.x) being used for command & control [2], but I don't know how that works because SK ISPs are extremely unlikely to permit any traffic to or from NK IP addresses. Remember, this is a country that censors virtually every NK-operated website. All in all, little direct evidence that links the attacks to NK has been released. Maybe such evidence exists but is classified. Maybe it doesn't exist in the first place. However, since the series of attacks over the last few years targeted SK's major financial, governmental, and media infrastructure, it is not too unreasonable to suspect that the attackers (or whoever hired them) have an axe to grind with SK as a whole, not with any particular corporation. Given the extremely polarized political landscape, it is also not unreasonable to suspect that someone who has a problem with SK might sympathize with NK. What doesn't get mentioned amid all the finger-pointing and political rhetoric is that SK's internet infrastructure desperately needs a major overhaul. The so-called "security" companies, who sell ActiveX controls with various alleged benefits, have been in charge of IT policy for over a decade. This has led to an online landscape where it is considered normal for any random website to ask the user to restart IE as administrator and install 5 ActiveX controls just to be able to log in. This is a completely separate problem from antiviruses. I'm not joking when I tell you that every bank requires you to install an ActiveX control that claims to detect and disable keyloggers. That's 1000x more evil than Norton Antivirus. [1] http://news.khan.co.kr/kh_news/khan_art_view.html?artid=201304102221475&code=940202 http://news.khan.co.kr/kh_news/khan_art_view.html?artid=2013... [Korean] [2] http://www.hani.co.kr/arti/economy/it/582269.html http://www.hani.co.kr/arti/economy/it/582269.html [Korean] [3] http://www.hani.co.kr/arti/economy/it/579266.html http://www.hani.co.kr/arti/economy/it/579266.html [Korean] [4] http://news.mt.co.kr/mtview.php?no=2013040922225524474&VRF http://news.mt.co.kr/mtview.php?no=2013040922225524474&V... [Korean] UPDATE: More links. ADDENDUM: In a previous comment [5], I argued that Ahnlab is an upstanding company that is unlikely to be involved in malicious activities. Turns out one of their products actually was exploited during the latest attack, though it was not their flagship antivirus suite. [5] https://news.ycombinator.com/item?id=5407719 https://news.ycombinator.com/item?id=5407719
- kyllo 14y ago
  Thanks for the summary. From this, it does appear that they used an antivirus update server as the attack vector. And about SK's internet infrastructure--believe me, I know. I had an account at Woori Bank when I lived in Korea. For internet banking authentication, they gave me two different PIN numbers (one 4-digit and one 6-digit) and a scratch-off card with a password matrix on it (at each login it would require me to enter digits found in different positions on the card), in addition to username/password and requiring me to install an AhnLab antivirus program, an "anti-keylogger" program, and 2-3 ActiveX encryption programs. I believe it was actually written in South Korean law until very recently that browser security/encryption MUST be implemented in ActiveX, and HTTPS/SSL was completely unheard of in Korea. I think these other authentication measures were devised to make up for the weaknesses of the ActiveX encryption. When I asked the bank rep why they had so many different authentication measures, the response was, and I quote, "Because Chinese people do the fake things."
  - kijin 14y ago
    The scratch-off card is a poor man's one-time password generator. In fact, it's remarkably effective at preventing scammers from draining people's accounts, as long as you're not stupid enough to take a photo of your card and post it online (yeah, people do that). That little card is what prevents the whole ActiveX fiasco from crumbling down like a house of cards, precisely because it's so low-tech (i.e. requires physical access). A lot of the other crap is understandable when you take into account the fact that online banking was first introduced to Korea in the mid-1990s. HTTPS was useless at the time because of U.S. export restrictions limiting it to 40 bits. Home-brewed ActiveX controls were used to implement 128-bit encryption. Korea also decided to use public keys and X.509 client certificates to identify people, way ahead of the rest of the world, but of course no browser supported them so someone had to write custom software. After that, inertia and regulatory capture allowed what was supposed to be a stopgap measure to continue for another 15+ years, long after proper HTTPS became available in every browser. The only benefit I can think of is that Koreans don't have to worry about those stupid password restrictions that a lot of American banks impose on their customers :)
techinsidr 14y ago
A few pieces here... http://www.securityweek.com/south-korea-cyber-attacks-used-data-wiping-trojan-component-wipe-linux-machines http://www.securityweek.com/south-korea-cyber-attacks-used-d... http://blog.trendmicro.com/trendlabs-security-intelligence/summary-of-march-20-korea-mbr-wiper http://blog.trendmicro.com/trendlabs-security-intelligence/s...

by kyllo 14y ago

drzaiusapelord 14y ago

Reason #452 why I never run AV on servers unless its a file server. AV is just a vector for either fat fingered definitions that destroy your system or a venue for malware to enter your system. I don't trust Sophos and Kaspersky that much. On the desktop, its a tolerated and perhaps necessary evil, but on servers - no way.

kijin 14y ago

My current understanding is that the malware was delivered as part of an update to some commonly used ActiveX controls. The initial infiltration took place 8 months ago, but the malware was designed to go live at a specific time [1]. Over 70 different pieces of malware were used to infiltrate a wide range of targets. The bash script was apparently used on non-Windows targets, of which there were only a few. EDIT: The single most effective attack vector seems to have been the update server for an ActiveX control called XecureWeb, which is used by several banks and their 20 million customers [4]. Just checked and I have XecureWeb too, in a VM that I only use for online banking. Good thing I have that VM turned off 99% of the time. According to SK news media, the primary reason the SK government suspects NK of the latest attack is because the attackers used some of the same malware, same proxy servers, and same techniques that were used in previous attacks, which the SK government also attributes to NK [2]. But AFAIK there is no indisputable evidence that links those previous attacks to NK, either. So there's a bit of circular reasoning involved in the accusation. The initial accusation of a Chinese attack was based on a Chinese IP address (101.106.25.105) that was used by the attackers. This accusation was rescinded when it was found that one of the banks that were hacked have been using the same IP range in its internal network </facepalm> [3]. But even if this wasn't the case, the case for a Chinese attack would be weak because anyone can use a Chinese open proxy. There was also mention of an North Korean IP address (175.45.178.x) being used for command & control [2], but I don't know how that works because SK ISPs are extremely unlikely to permit any traffic to or from NK IP addresses. Remember, this is a country that censors virtually every NK-operated website. All in all, little direct evidence that links the attacks to NK has been released. Maybe such evidence exists but is classified. Maybe it doesn't exist in the first place. However, since the series of attacks over the last few years targeted SK's major financial, governmental, and media infrastructure, it is not too unreasonable to suspect that the attackers (or whoever hired them) have an axe to grind with SK as a whole, not with any particular corporation. Given the extremely polarized political landscape, it is also not unreasonable to suspect that someone who has a problem with SK might sympathize with NK. What doesn't get mentioned amid all the finger-pointing and political rhetoric is that SK's internet infrastructure desperately needs a major overhaul. The so-called "security" companies, who sell ActiveX controls with various alleged benefits, have been in charge of IT policy for over a decade. This has led to an online landscape where it is considered normal for any random website to ask the user to restart IE as administrator and install 5 ActiveX controls just to be able to log in. This is a completely separate problem from antiviruses. I'm not joking when I tell you that every bank requires you to install an ActiveX control that claims to detect and disable keyloggers. That's 1000x more evil than Norton Antivirus. [1] http://news.khan.co.kr/kh_news/khan_art_view.html?artid=201304102221475&code=940202 http://news.khan.co.kr/kh_news/khan_art_view.html?artid=2013... [Korean] [2] http://www.hani.co.kr/arti/economy/it/582269.html http://www.hani.co.kr/arti/economy/it/582269.html [Korean] [3] http://www.hani.co.kr/arti/economy/it/579266.html http://www.hani.co.kr/arti/economy/it/579266.html [Korean] [4] http://news.mt.co.kr/mtview.php?no=2013040922225524474&VRF http://news.mt.co.kr/mtview.php?no=2013040922225524474&V... [Korean] UPDATE: More links. ADDENDUM: In a previous comment [5], I argued that Ahnlab is an upstanding company that is unlikely to be involved in malicious activities. Turns out one of their products actually was exploited during the latest attack, though it was not their flagship antivirus suite. [5] https://news.ycombinator.com/item?id=5407719 https://news.ycombinator.com/item?id=5407719

kyllo 14y ago
Thanks for the summary. From this, it does appear that they used an antivirus update server as the attack vector. And about SK's internet infrastructure--believe me, I know. I had an account at Woori Bank when I lived in Korea. For internet banking authentication, they gave me two different PIN numbers (one 4-digit and one 6-digit) and a scratch-off card with a password matrix on it (at each login it would require me to enter digits found in different positions on the card), in addition to username/password and requiring me to install an AhnLab antivirus program, an "anti-keylogger" program, and 2-3 ActiveX encryption programs. I believe it was actually written in South Korean law until very recently that browser security/encryption MUST be implemented in ActiveX, and HTTPS/SSL was completely unheard of in Korea. I think these other authentication measures were devised to make up for the weaknesses of the ActiveX encryption. When I asked the bank rep why they had so many different authentication measures, the response was, and I quote, "Because Chinese people do the fake things."
- kijin 14y ago
  The scratch-off card is a poor man's one-time password generator. In fact, it's remarkably effective at preventing scammers from draining people's accounts, as long as you're not stupid enough to take a photo of your card and post it online (yeah, people do that). That little card is what prevents the whole ActiveX fiasco from crumbling down like a house of cards, precisely because it's so low-tech (i.e. requires physical access). A lot of the other crap is understandable when you take into account the fact that online banking was first introduced to Korea in the mid-1990s. HTTPS was useless at the time because of U.S. export restrictions limiting it to 40 bits. Home-brewed ActiveX controls were used to implement 128-bit encryption. Korea also decided to use public keys and X.509 client certificates to identify people, way ahead of the rest of the world, but of course no browser supported them so someone had to write custom software. After that, inertia and regulatory capture allowed what was supposed to be a stopgap measure to continue for another 15+ years, long after proper HTTPS became available in every browser. The only benefit I can think of is that Koreans don't have to worry about those stupid password restrictions that a lot of American banks impose on their customers :)

techinsidr 14y ago

A few pieces here... http://www.securityweek.com/south-korea-cyber-attacks-used-data-wiping-trojan-component-wipe-linux-machines http://www.securityweek.com/south-korea-cyber-attacks-used-d... http://blog.trendmicro.com/trendlabs-security-intelligence/summary-of-march-20-korea-mbr-wiper http://blog.trendmicro.com/trendlabs-security-intelligence/s...