Technical Paper

Storage layers

1.4 Storage layers

The storage model separates data by purpose and privacy class. The public document describes which data lives in which layer.

LayerContentPlacementRetention principlePrivacy class
Hot record layerReceipt records, line items, stage eventsApplication databaseActive product windowPseudonymous
Analytics layerNormalized observations and quality metricsSeparate analytics partitionPolicy-defined rolling windowPseudonymous or anonymous
Object layerEncrypted receipt input and processing derivativesEncrypted object storeData-minimization policyMay contain personal data
Anonymous aggregate layerAggregate output for the B2B data productSeparate aggregate storeVersioned publication windowNot linkable back to user
On-chain summaryToken event, settlement commitment, program statePublic chainPermanentToken and commitment data

Two rules are invariant: raw receipt content is processed in the off-chain data layer; the anonymous aggregate layer uses user-separated keys. Retention periods and physical provider selection are operational policy.