Business Continuity Planning & Disaster Recovery

the family of statistical and machine-learning techniques (classification, clustering, association, regression, anomaly detection) that extract patterns from that warehouse.

is the combination of software, hardware, and telecommunication networks used to collect useful data, organise it, and turn it into actionable information for an enterprise.

Purpose: Process day-to-day transactions; Time-horizon: Current — seconds, minutes; Read/Write profile: Write-heavy, short queries; Example in a bank: Finacle/Flexcube CBS, UPI switch, NEFT/RTGS engine, ATM switch

Purpose: Analyse data and support decisions; Time-horizon: Historical — months, years; Read/Write profile: Read-heavy, long aggregate queries; Example in a bank: Enterprise Data Warehouse (EDW), MIS, ALCO dashboards, BCBS 239 risk-data hub

is a type of data management system purpose-built to enable and support business activities, especially analytics.

Inmon School (Top-Down): First the EDW (single enterprise warehouse in 3NF), then carve data marts out of it; Kimball School (Bottom-Up): First the data marts (one per business process, dimensional star schemas), then federate via conformed dimensions into the EDW

Inmon School (Top-Down): 3NF / normalised (corporate information factory); Kimball School (Bottom-Up): Star / snowflake (dimensional, denormalised for analytics)

Inmon School (Top-Down): Long (12-24 months for the EDW); Kimball School (Bottom-Up): Short (3-6 months per data mart)