What's Ahead for Big Data?
summary
Summarizes where Hadoop/MapReduce and NoSQL succeed and fail, gives concrete alternatives (Cascading/Scalding, Spark over MapReduce, Hive/Impala/Presto), explains when to use distributed filesystems vs. databases (Cassandra, Titan for graphs), and covers streaming (Storm, Spark streaming) plus ML tool choices (Mahout, Spark ML). States performance gaps (disk flushes, batch latency) and practical tradeoffs with examples and claimed speedups, so decide by which tool and workload match.
AI-written summary · about
chapters (11)
- 0:00Start
- 0:07What's Ahead for Big Data?
- 4:52Formal Schemas
- 7:20Big Data Architectures
- 12:32A Hadoop Cluster
- 12:53MapReduce in Hadoop
- 13:38Compute Inverse Index
- 14:49Anatomy: MapReduce Job
- 17:54MapReduce is designed for offline, batch-mode
- 41:17Open-source Alternatives
- 45:20MapReduce is waning
related talks
45mFrom Big Data to Big Information
15mOptimizing Query Performance in Big Data Systems
39mBig Data Systems Performance: The Little Shop of Horrors
27mNavigating Unstructured Data - Availability vs Analytics in NoSQL
42mContemporary Approaches to Data at Scale
01h 01mNoSQL: An Analysis
44mReal World Use Cases Hadoop and NoSQL in Production
08mHadoop and Big Data Hype