Hadoop Security: Seven Ways to Kill an Elephant
summary
Claim: Hadoop distributions ship insecure defaults and show concrete failures — Kerberos is treated as a flip‑switch so teams issue long‑lived tokens, delegation tokens appear in URLs and many JDBC/Hive connections run without SSL or with null ciphers. Tools clash: Scoop2 disables security mode; Storm instances require rm -rf restarts. Real incidents include an intern’s DROP TABLE that forces a 96‑hour restore. Proposed fixes: central service layer, per‑query authorization, and GRC layered over accumulo/Tahoe proxies.
AI-written summary · about
chapters (54)
- 0:00Introduction
- 2:54What is your job
- 3:47The Sam Jackson Audience Dilemma
- 5:40The genesis of Hadoop
- 8:46The future of technology
- 14:17Big data analysis
- 16:21Big data and health care
- 17:39Algorithmic accountability
- 18:12Audience operating system
- 18:59Quiz
- 20:48The scary part
- 21:36The rearview mirror
- 22:04Race cars
- 22:41Seatbelts
- 23:03Regulators
- 24:40Hadoop Ecosystem
- 25:47Scoop
- 26:16Storm
- 27:47Overview
- 28:17My Argument
- 28:32Zombie vs Terrorist
- 29:32God Mode
- 30:08Go Game
- 31:51Hadoop Security
- 34:20Blade Runner
- 35:29Monitoring
- 36:15Lines in Gravel
- 37:47Flume
- 38:08StormWrite
- 40:28Pivotal
- 40:55Command Center
- 43:05Authentication Perimeters
- 44:34Domain Perimeter
- 45:26Vulnerability Management
- 45:58Delegation Tokens
- 46:47JIRA
- 47:20Cow
- 47:51Cholera
- 48:22ERM
- 48:52HTTP Services
- 49:26MapReduce
- 49:53Public Information
- 50:08My Brother
- 50:20We Dont Care
- 51:40Data Protection is Weird
- 52:05Auto Autos
- 52:36Big Data
- 53:48Network System Solution
- 54:13Why Does That Matter
- 55:01Example
- 56:28Future State
- 56:59Compliance Standards
- 57:30Data Services
- 59:35Recap
related talks
50mBabar-ians at the Gate: Data Protection at Massive Scale
10mAdding Security to Hadoop
31mRings of Security for Big Data in Cloud
15mThe Future of Hadoop: What Happened & What's Possible?
03mApache Hadoop Meets Cybersecurity
34mAvoiding Big Data Antipatterns
01h 29mHadoop 201 -- Deeper into the Elephant
45mDeep Dive into Data Streaming Security
