Point One Zero Platform architecturev11 · generated viewSign in to edit
Platform Layers › L6 Data & Documents
L6 Data & Documents
Where this layer sits on the architecture; the client plane band to its right shows the values it reads.
L6 Data & Documents PER-TENANT ISOLATION · DATA ENGINEERING
The storage half, split from Integration because the technology and the skillset differ (data engineers and architects: schema, indexing, retrieval quality). The stores are reusable templates; the client's instances and data are plane-side.
Vector DB, document store, SQL/warehouse services + retrieval APIs, provisioned per tenant from templates
Indexing & embedding jobs; taxonomy enforcement (document structures come from Knowledge)
Plane-side: tenant store instances (the client's actual data), document-to-taxonomy mapping, residency, retention & legal hold
Isolation verified by automated cross-tenant tests in CI
Restricted stores stay walled: excluded from every shared index and pipeline; entitlement-checked retrieval only
Never the system of record: originals stay in the client's source estate, every store here holds derived, disposable copies
Boundary test: nothing in another layer stores client data; nothing here knows what the data means.
Client-specific configuration for this layer is specified band by band in Client Delivery, C6 spec: every input with its binding key, generated artifact, and verifying check.
ImplementationProvisioned from templates at setup; per-engagement case indexes follow the scoped index lifecycle: created at open, destroyed at close.
Document Store Service
The platform's working copies of documents, with lineage attached.
ImplementationDerived, disposable copies only: the originals remain in the client's source estate, which stays the system of record.
Data Warehouse Service
Structured analytical storage per tenant.
ImplementationProvisioned per tenant from templates, or connected to the client's existing warehouse through the Integration layer instead of migrating it.
Retrieval APIs
The only read path agents use into tenant data.
ImplementationEntitlement-checked on every call; restricted stores are reachable only here; responses follow the high-signal, paginated tool standard.
Indexing & Taxonomy Enforcement
Classification and embedding jobs that keep stored content aligned to Knowledge's document taxonomy.
ImplementationStructures come from Knowledge; execution belongs to platform machinery in L7; taxonomy mapping per client is a C6 value.
Systems of Record
The client's ERP, CRM and fund administration systems.
ImplementationNever replaced and never written around: reached through the south servers, with the authority mapping deciding which source wins a disagreement.
Document Sources
VDR-class rooms, SharePoint, network drives: where documents actually live.
ImplementationRestricted stores stay walled: excluded from every shared index and pipeline, entitlement-checked retrieval only.
Unstructured Data
Notes, transcripts, recordings and similar loosely structured material.
ImplementationEnters through the ingestion pattern (P3) with PII screening at the door, before anything is stored or indexed.
External Data Providers
Market and benchmark data feeds.
ImplementationReached through the API gateway; programmatic-access license terms are captured in Discovery section D before any connection is built.