How BT Uses Hadoop with its CRM
summary
BT builds a multi-tenant Hadoop platform (Cloudera-based) to replace a daily ETL CMF process comparing nearly a billion records across 12 systems; first production use case is master data management for BT UK business customers. A 50-node cluster running mapreduce Java cuts the cycle from 24 to 8 hours (15x velocity), handles 5x the data and yields ~200–250% ROI. Pipeline work includes inventory-backed ultrafast broadband qualification, TV-to-network correlation, millions of 4G-in-home telemetry and local IoT (MK:Smart) integrations.
AI-written summary · about
chapters (7)
- 0:00Intro
- 0:13Our purpose, goal, strategy and culture
- 1:36Customer Expectations
- 2:23BT Consumer Customers
- 5:36Customer Master File ETL Pipeline
- 7:59Projects in development
- 11:04Data Architecture Vision
related talks
57mBig Data: Threat or Opportunity?
23mDelivering Modern Data Stacks With dbt to Operationalise Analytics
52mUsing Hadoop at Badoo
42mBillion-Scale Triumph: Innovating at the Edge of Cloud Data
52mUsing Hadoop at Badoo
05mHadoop in 5 Minutes or Less
10mSeparating Hadoop Myths from Reality
47mSolving Real & Big (Data) Problems Using Hadoop
