Clojure Versus the Botnets
summary
Clojure becomes Damballa’s production language for ETL, modeling and microservices, enabling rapid malware‑detection ML pipelines that run on half a terabyte a day of ingest and about a quarter terabyte a day of passive DNS for the Penumbra project. Code integrates directly with Hadoop, Avro, Parkour, HyperLogLog, LIBLINEAR (Java port) and small Mahout pieces, uses REPL-driven iteration and pure value transforms, and yields classifiers for botnet/click‑fraud behavior that were built quickly and operated at Internet scale.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:19How do computers work
- 1:01Why Clojure
- 1:35What we do
- 2:05What we use
- 2:46Data lake
- 3:47Data like architecture
- 5:43Penumbra
- 6:44JVMHosted Pneus
- 9:36Intersections
- 11:36Value Orientation
- 13:31Transformations
- 16:53Conclusion
related talks
44mData Science in Clojure
37mClojure, a Sweetspot for Analytics
29mDeep Learning Needs Clojure
40mBuilding Machine Learning Models with Clojure and Cortex
33mClojure Data Science
48mCatch me if you can: Building a Web Malware Analyzer using Machine Learning
44mApplied Machine Learning for Data Exfil and Other Fun Topics
41mDeploying ML models in a Clojure environment