Skip to content

Threat Analysis of Complex Attacks (SANS DFIR 2015)

Talk on YouTube

You can't prevent complex attacks

  • When infected by a cyber criminal's malware, your job is just to get rid of the malware
  • But for complex attackers, malware is just a tool. They can also hire people to infiltrate you, for example
  • You can't prevent complex attacks

Don't let APTs know once you detect them

  • Don't submit to VirusTotal
  • Don't publish Yara rules

If you do, they'll just create new malware and return. Remember, these are targeted attacks

Instead...

  • Limit their bandwidth
  • Feed them false information
  • This will buy you time to investigate the attacker

A few ways to detect APTs

Outgoing network traffic

  • store DNS queries your endpoints make (use OpenDNS)
  • capture full packet captures with Suricata
  • why outgoing? malware must talk to C&C server over network
  • I wonder if malware could be controlled by a rogue employee regularly plugging in a USB drive

but APTs try bypassing this!

  • Equation Group sends its outgoing traffic (responses to C&C server?) to US universities
  • these universities then somehow send the packets to Equation Group

Honey docs

  • Spin up a server, add documents to it, send alert whenever a doc is opened

Technical analysis of samples

  • APTs reuse code samples; after all, they spend a lot of money developing them
  • You should be able to recognize attacker's code
  • Have an internal threat intel tool
  • For each bad IP or domain you see, write down who you think it is, all the times you saw them, etc
  • Potentially, 2 threat actors will work together, or really be 2 campaigns of one threat actor. Detect this if 2 threat actors are using the same IPs

Mapping victims

Using VirusTotal

  1. Write YARA rule
  2. Setup VirusTotal to email you whenever a piece of malware, which matches your rule, is uploaded

Using domains

  • Register an domain name the APT used in malware
  • Can find domain names hard-coded in malware
  • Then, you'll get traffic from all the victims

Attribution

  • Attribution is easy to fake!
  • Try recovering the passwords of the C2 server
  • May be hard-coded in the malware
  • Can then log into C2 server. If malware creator forgets to use a VPN, you can see what IP he has
  • Have a linguistic expert on the team
  • He said a European person probably misspelled "exceeded" but that's not convincing
  • Having a geopolitical expert could be useful to understand the attacker's motivations

Protection

Austrialian Signals Directorate: To stop 85% of complex attacks, do:

  1. Whitelisting executables and default deny
  2. Patching apps
  3. Patching OS
  4. Restrict admin privileges

Who are the threat actors?

  • Insider threats - motivated by money or revenge
  • Hacktivist networks - motivated by ideological reasons
  • APTs - motivated by economic, military reasons
  • Script Kiddies - motivated by reputation
  • Cyber criminals - motivated by money
  • Rogue hackers - motivated by reputation

Building a cyber army for North Korea

Talk on YouTube

Strength in cyberspace = controlling as many machines as possible

  • Makes attributing to you hard
  • And attributing to your opponent easy. (If you are on a network which your opponent hacks)
  • Makes denial-of-service attacks feasible