Designed for Security - IoT Big Data End to End Best Practices

IoT Tech Expo - Europe
10 -11th Feb, Olympia, London, 2016

http://www.iottechexpo.com/europe/track/data-security/

- Challenges in IoT and security, privacy
- IoT = Big Data
- Best practices of IoT Big Data Modelling
- Latest development in Guaranteed Privacy and Secure MPC (multi-party computing)

Transcript

1. Designed for Security - IoT Big Data End to End Best Practices • Challenges in IoT and security, privacy • IoT = Big Data • Best practices of IoT Big Data Modelling • Latest development in Guaranteed Privacy and Secure MPC (multi-party computing) 10-11 February 2016, Olympia London #IoTTechExpo Charles Cai, Chief Architect | Senior Advisor, Finance Group | Deputy GM, IDC| Wanda
2. Designed for Security – IoT Big Data Modelling • #FO #FICC: Investment Banking Front Office: FX/Commodities • #ETRM: Energy Trading & Risk Management • #entrepreneur #innovator #disruptor • Hacker / Maker: won major IoT / FinTech / BlockChain/ BigData hackathons; Organizer of high profile big data hackathons • Twitter: @caidong #big-data #data-science #IoT #cloud • LinkedIn: http://uk.linkedin.com/in/charlescai/en • Email - business charlescai@wanda.cn • Email - IoT: charles.cai@socialogix.net 10-11 February 2016, Olympia London #IoTTechExpo Charles Cai
3. Norse Attack Map
4. “There's now a growing sense of fatalism: It's no longer if or when you get hacked, but the assumption is that you've already been hacked, with a focus on minimizing the damage.” Source: Dark Reading / Security’s New Reality: Assume The Worst
5. Breaches Happen in Hours… But Go Undetected for Months or Even Years Source: 2013 Data Breach Investigations Report Seconds Minute s Hours Days Weeks Months Years Initial Attack to Initial Compromise 10% 75% 12% 2% 0% 1% 1% Initial Compromise to Data Exfiltration 8% 38% 14% 25% 8% 8% 0% Initial Compromise to Discovery 0% 0% 2% 13% 29% 54% 2% Discovery to Containment/ Restoration 0% 1% 9% 32% 38% 17% 4% Timespan of events by percent of breaches In 60% of breaches, data is stolen in hours 54% of breaches are not discovered for months
6. https://www.youtube.com/watch?v=MK0SrxBC1xs http://blog.atmel.com/2015/05/04/the-10- challenges-of-securing-iot-communications/ IoT can easily go wrong …
7. Hardware Renaissance started with Open Source Hardware
8. From M2M to IoT, to Cloud to Big Data
9. IoT OSI Stack – Where to look for threat surface
10. Case Study: Continuous Patient Monitoring with Wearables + IMU + bio-sensors • Until now there’s no cure for Parkinson’s disease • New medicine trial is an extremely slow process – and feedbacks from the patients are not frequent at all • Wearables + IMU + BLE can easily help quicken the process • Sensor data can be collected from Microsoft Band in real-time (up to 64 measurements per second): • Movements: gyroscope, accelerometer, digital compass, … • Bio-metrics: heartbeat, skin temperature, … • To analyze typical Parkinson’s symptoms: tremor and slowed movements – Loss of automatic movements, impaired posture and balance • Along with clinical records of medicine intake • Technology used: Wearables with IMU + BLE -> IoT Gateway -> Cloud
11. Case Study: Advanced Healthcare Machine Learning / Predictive Platform in the Cloud • By collecting continuous bio-sensor data, we need big storage + advanced machine learning model • Cloud based scalable time series database + Streaming Analytics + Machine Learning Stack provides a sophisticated platform to tackle the big data challenge in Healthcare • 9-axis IMU (Inertial Measurement Unit) with Gyroscope + Accelerometer + Digital Compass • Advanced sensor fusion to be developed • Classification of wearer activities • sitting, standing, walking, running, sleeping… • Detect pattern of patient symptoms • E.g.: Parkinson’s Disease: • predicting deterioration speed • new trial medicine effectiveness • 20+ other environmental + biometric sensor data • Easily 200MB – 10GB uncompressed data a day!
12. Home Environmental Sensors Bio-sensing Wearable Devices Personal Bio-metrics Health tracking Speech, face and emotions recognition, interaction Home Environmental Metrics Predictive Home Care Analytics dslogix Internet of Things Reminders, tasks monitor movement tracking Doctors / Carers / MIners Family Other IOT Other Patients Intelligent Algorithms Big Data IoT Gateway w/ Connected e-Health Instruments Smart Medicine Container Smart Cane https://dsrobotix.io e-Diagnostics VR / AR Case Study: End-to-End Connected Health
13. Low Power = Low Security! The key is IoT Gateway • AES Computation • BLE 4.1 • CoAP vs JSON • HTTPS / TLS • ARM Cortex – M0, *M3/M4… Some vendors: • Intel/Dell IoT Gateway • Cryptosoft IoT Security Solution • …
14. So IoT = Big Data - Where are we at with Big Data Analytics? By Thomas Davenport – Harvard Business Review
15. IoT + Big Data / Data Science for all industries
16. Open Source Data Science Toolbox Hadoop / Mesos Distributed Storage + Scalable Computation 16 COTS Apps (Excel, Tableau, Qlik...) Statistical Time Series Analysis Wider Big Data Analytics eco-systems • Shell/APIs: HDFS, Hive, Spark, HBase, Sqoop, JDBC/ODBC • Languages: Julia, Python, R, Scala - Developed on: - Operated by: NLTK: Natural Language Distributed Time Series / Geospatial / Graph Databases GIT Repo D a t a P r o d u c t s WebSocket Drag + Drop (CZML/GeoJSON) Web Browser (collaboration) Export to CSV/Excel Geospatial data Time Series data Public Data Market data Real-time Streaming Open Gov Data JDBC via phoenix HDFS Hive/Pig w/ Geospatial Open Source Enterprise Big Data / Data Science Platform Attack Points
18. AAAA & Hadoop Security Best Practices
19. Introducing OpenSOC Intersection of Big Data and Security Analytics Multi Petabyte Storage Interactive Query Real-Time Search Scalable Stream Processing Unstructured Data Data Access Control Scalable Compute OpenSOC Real-Time Alerts Anomaly Detection Data Correlation Rules and Reports Predictive Modeling UI and Applications Big Data Platform Hadoop
20. Simple data masking is not enough! • New York “Anonymous” Taxi Data Privacy Leak • Consider: • K-anonymity • l-diveristy • T-closeness • Q-disclosure • Q-precense PII (Personal Identifiable Information) needs advanced de- identification / anonymous process
21. New Development: Secure Multi-party Computation Wikipedia Homomorphic Encryption –> Computation while Encrypted Multi-party Computation -> Privacy Preserved Computation
22. Test: How to run a simple secured MapReduce Test: Apache Spark with BlockChain integrated
23. Benefits • Legal requirements: e.g. • Privacy-preserving • Auditable – who can access / compute my data • Full + pseudo-anonymous (k-anonymity) • Monetizing opportunity for data contributors (based on computation counters) • Kill-switch: control if he/she wants to participate in the computation • Meeting HIPPA: • Ethereum VM, Esri DB, Solidity (need extension in Smart Contract for private data, a.k.a. Private Contract) + Spark Analytics Computation Engine
24. Summary 1. Holistic security design from low powered wearables to On-Prem / Cloud based Big Data Analytics 2. IoT = Big Data 3. Big Data = Big Responsibility 4. Privacy Preserved Computation +