Jim Campigli - BigDataNYC 2014 - theCUBE - #BigDataNYC
summary
Claims a patent-backed active-active replication for HDFS that runs multiple active name nodes across data centers, giving full read/write synchronous copies and continuous hot backup so no idle standby hardware is needed. Enables multi-data-center ingest and a single global namespace, enforces consistency automatically to avoid regulator-triggering drift, and adds cluster zoning to route hot workloads to high-spec servers. Demos an alpha unification layer for Hortonworks, Cloudera and MapR.
AI-written summary · about
chapters (16)
- 0:00Introduction to BigDataNYC
- 0:41Welcome to Jim Campigli
- 1:44Why Hadoop in Financial Services
- 2:47Extracting value from PDFs with Hadoop
- 3:52Active active 24x7 resiliency
- 5:39Multi-data center ingest benefits
- 7:25Is the data mover obsolete?
- 8:52Customer maturity with Hadoop
- 10:37Competition in active active solutions
- 13:02Eliminating planned downtime
- 14:07IT infrastructure priorities
- 14:51Cluster zoning explained
- 17:59Future use cases and growth
- 19:04Hadoop distribution integration
- 21:10IoT and mission critical demand
- 23:37Closing remarks and next event
