Presentation is loading. Please wait.

Presentation is loading. Please wait.

Fast and Precise In-Browser JavaScript Malware Detection

Similar presentations

Presentation on theme: "Fast and Precise In-Browser JavaScript Malware Detection"— Presentation transcript:

1 Fast and Precise In-Browser JavaScript Malware Detection
ZOZZLE: Fast and Precise In-Browser JavaScript Malware Detection

2 What is the Problem? JavaScript allows authors to run any code when a user visits a web page JS-based malware attacks are the majority of successful mass-scale exploitation Malware is easy to hide: self-generating code that produces more code to run JS severs important functionality for many sites In-browser solutions have not been fully accepted because of the performance hit Browsers use offline scanning to check URLs but there are too many sites and malware typically comes and goes frequently

3 Challenges Performance Accuracy Obfuscated malware Malware transience
Detection is not fast enough to be used in a browser Accuracy False positive rates of 5% is acceptable for static analysis tools but is over 100x what is acceptable for in-browser detection Obfuscated malware Most JavaScript code is frequently obfuscated so purely static detection is generally ineffective Ex. eval, document.write generate code at runtime that is difficult to pattern-match Malware transience Offline-only scanning is not effective because web malware “infects fast and dies young” Nearly 20% of malicious URLs were gone after 1 day

4 To increase the changes to successful exploitation, multiple exploits often exist within the same page eval, <iframe>, <script> unfolding reveals obfuscated code, but depth is not a good indicator Used by JavaScript libraries to save space through client-side code generation Used as weak copy protection to avoid code piracy

5 Solution : Zozzle Performance Accuracy De-obfuscation
AST-based detection is fast and scalable Fast classification: throughput at over 1 MB of JavaScript code per second Accuracy AST-based detection uses hierarchical (context-sensitive) features more precise than text-based Low false positive rate: % (< 1 in 1/4 million) De-obfuscation Uses JavaScript engine of a browser to expose obfuscation and get the final, expanded version of JavaScript code

6 What Is Zozzle? A highly precise, mostly static detector for malware written in JavaScript suitable for in- browser deployment 3 Steps: JavaScript context collection and labeling as benign or malicious Feature extraction and training of a naïve Bayesian classifier Applying the classifier to a new JavaScript context to determine if it is benign or malicious

7 Zozzle: How It Works JavaScript runtime engine exposes attempts to obscure malware JS code is unfolded to just before it’s executed Intercept calls to compile() in the JavaScript engine It’s invoked when eval is called and whenever new code is included with an <iframe> or <script> tag Observe JS code at each level of its unpacking just before it's executed by the engine.

8 How It Works cont. A static classifier trained with a context-sensitive AST (abstract syntax tree) and a collection of labeled malware samples analyzes JS Nozzle runtime detector dynamically crawls millions of URLs and collects sample malware by observing the behavior of running JS code Tries to avoid transience and cloaking by scanning a wide range of URLs

9 Benign vs. Malicious Samples

Download ppt "Fast and Precise In-Browser JavaScript Malware Detection"

Similar presentations

Ads by Google