1

Introduction

• The origins of Big Data: A world of digital data, e-health, timeline.
• The four-V's definition: Origins of the data.
• A breakthrough: Changes in quantity, quality, and habits.
• The value of data: A change in importance.
• Data as a raw material.
• The fourth paradigm of scientific discovery.

2

Big Data: Processing, from acquisition to result.

• The sequence of operations. Acquisition.
• Data collection: crawling, scraping.
• Managing event flows (Complex Event Processing, CEP).
• Indexing incoming flows.
• Integration with old data.
• Data quality: A fifth V?
• Different types of processing: Searching, learning (Machine Learning, transactional learning, data mining).
• Other sequencing models: Amazon, e-Health.
• One or more data repositories? From Hadoop to the in-memory.
• From tonal analysis to knowledge discovery.

3

Relationships between the Cloud and Big Data

• The architecture model of public and private Clouds.
• XaaS services.
• The goals and benefits of Cloud architectures.
• Infrastructure.
• Similarities and differences between the Cloud and Big Data.
• Storage clouds.
• Classification, security, and privacy of data.
• Structure as a classification criterion: Unstructured, structured, semi-structured.
• Classification by life cycle: Temporary or permanent data, active archives.
• Security difficulties: Increased volumes, distribution.
• Potential solutions.

4

Introduction to Open Data

• Philosophy of open data and goals.
• Releasing public data.
• Implementation difficulties.
• Essential features of open data.
• Areas involved. Expected benefits.

5

Equipment for storage architectures

• Servers, disks, networks, and use of SSD drives, importance of network infrastructure.
• Cloud architectures and more traditional architectures.
• Benefits and difficulties.
• The TCO. Power consumption: Servers (IPNM), drives (MAID).
• Object storage: principle and benefits.
• Object storage compared to traditional NAS and SAN storage.
• Software architecture.
• Storage management location levels.
• Software-Defined Storage.
• Centralized architecture (Hadoop File System).
• Peer-to-peer and hybrid architectures.
• Interfaces and connectors: S3, CDMI, FUSE, etc.
• Future of other storage types (NAS, SAN) relative to object storage.

6

Data protection

• Preservation over time in the face of increased volumes.
• Online or local backups?
• Traditional archiving and active archiving.
• Links with storage hierarchy management: Future of magnetic tape.
• Multisite replication.
• Damage to storage media.

7

Scope processing methods

• Classification of analysis methods based on data volume and processing power.
• Hadoop: The Map Reduce processing model.
• The Hadoop ecosystem: Hive, Pig. The difficulties of Hadoop.
• OpenStack and the Ceph data manager.
• Complex Event Processing: An example? Storm.
• From BI to Big Data.
• Return to decisional and transactional models: NoSQL databases. Types and examples.
• Data ingestion and indexing. Two examples: Splunk and Logstash.
• Open-Source crawlers.
• Search and analysis: Elasticsearch.
• Learning: Mahout. In-memory.
• Visualization: Real-time or not, in the Cloud (Bime), comparison of QlikView, Tibco Spotfire, and Tableau.
• A general architecture of data mining via Big Data.

8

Usage case through examples and conclusion

• Anticipation: Needs of users within companies, equipment maintenance.
• Security: People, fraud detection (mail, taxes), the network.
• Recommendation. Marketing analysis and impact analyses.
• Path analyses. Distribution of video content.
• Big Data for the automotive industry? For the oil industry?
• Should you begin a Big Data project?
• What future is there for data?
• Governance of data storage: Roles and recommendations, Data Scientists, skills involved in a Big Data project.