Rishi Yadav - BigDataSV 2014 - theCUBE
summary
Consultancy builds customer IP on OpenSource Hadoop, using Yarn to run Storm, Spark and Giraph natively so real‑time and graph apps live where the data is; claim that Spark replaces MapReduce by operating in memory and that slave nodes often have 32–48 GB RAM each. Drill and Impala demonstrate low‑latency SQL on Hadoop, visualization work uses d3.js visualizations, and mid‑market clients adopt Hadoop now for production use.
AI-written summary · about
chapters (18)
- 0:00Welcome to BigDataSV 2014
- 1:25Honest opinions on Big Data
- 2:08Pretenders vs winners in Hadoop
- 3:11Data lakes and YARN revolution
- 4:57YARN’s role in Hadoop evolution
- 6:44InfoObjects consulting focus
- 8:06Next-gen analytics and visualization
- 9:27Real-time data trends
- 11:12Big Data developer skills
- 12:58Programming languages for Big Data
- 14:22Hyperscale trends and Big Data
- 16:29Big Data as decade-defining trend
- 17:10Competing with Big Data giants
- 17:51SMBs adopting Big Data early
- 19:14ROI and skepticism in 2014
- 20:18People as data and social trends
- 21:01Hadoop marketplace growth
- 22:06BigDataSV 2014 summary
