Brian Bulkowski - BigDataSV 2014 - theCUBE
summary
Announces a new node.js client released today and an open source recommendation engine built on non‑contextual cosine distance/vector approaches, meant to run with high IOPS and random access on SSDs. Claims a terabyte of DRAM costs about $1M versus flash at a few thousand dollars, and points to a reported $1.2B uplift from real‑time personalization. Offers a click‑to‑spin two-node flash cluster via Internap so teams can prototype without installing anything.
AI-written summary · about
chapters (11)
- 0:00Introduction to BigDataSV 2014
- 0:45Aerospike's in-memory focus
- 3:11Expanding to mainstream adoption
- 5:35Real-time data's business impact
- 6:58Integrating real-time with Hadoop
- 9:05Overcoming legacy system barriers
- 9:46Polyglot databases and SQL
- 10:49Open source and enterprise trust
- 12:56Media and advertising success
- 14:19Developer experience and tooling
- 15:20Aerospike's value proposition
