Robert Hodges | Hadoop Summit 2014
summary
Continuent describes a working pipeline that does real-time replication from MySQL and Oracle into Hadoop using log‑based change data capture, designed to add low load and no application changes. Production footprint: 100 customers, hundreds of millions of transactions daily and examples of a customer with ~8 billion transactions a day across ~50 systems and single OLTP stores of 7–8 TB. Latency goals range from ~5 minutes to 24 hours and the replication code is open source.
AI-written summary · about
chapters (14)
- 0:00Introduction to Hadoop Summit
- 1:06Company background and focus
- 2:07Structured vs unstructured data in enterprises
- 3:12Relational databases to Hadoop migration
- 4:34Real-time data replication challenges
- 6:59Key requirements for enterprise solutions
- 8:00Maturity of the Hadoop ecosystem
- 9:47Open source innovation in Hadoop
- 10:28Competitive landscape in data movement
- 11:52Importance of real-time data processing
- 12:56Big data and cloud computing collision
- 14:44Hadoop vs cloud computing models
- 15:05Industry inflection point in data storage
- 16:30Future plans for the company
