Multi‑Tenant Architecture

Overview

ClearLedger Analytics is designed as a local, multi‑tenant engine for regulated portfolio teams. Each tenant — firm, desk, or advisor group — operates inside its own sealed environment, with isolated data, configuration, and optimization workflows. There is no shared cloud state, no cross‑tenant data leakage, and no telemetry streaming out of the engine.

This architecture combines ClearLedger’s local, in‑memory VertiPaq engine with Hardened Engine Security to deliver a multi‑tenant model that is both operationally efficient and audit‑friendly.

1. Tenant Isolation Model

In ClearLedger, a “tenant” is a logically isolated environment that contains:

its own history files and LocalMaster metadata
its own optimization constraints and configuration
its own ML cache and signal outputs
its own portfolio definitions and export paths

Tenants do not share:

position data
risk structures
ML signals
optimizer state

Isolation is enforced at the file‑system and engine level, with each tenant mapped to its own directory structure and environment configuration.

2. Local, In‑Memory Execution Per Tenant

Each tenant runs on a local, in‑memory VertiPaq engine

This provides:

deterministic analytics per tenant

no cross‑tenant performance interference

no shared cloud compute or external APIs

consistent behavior across sessions and machines

Excel acts as the UI shell, while the engine executes all calculations inside the tenant’s sealed context.

3. Hardened Engine Security for Multi‑Tenant Use

ClearLedger’s hardened engine design is critical for multi‑tenant deployments:

Zero telemetry — no outbound analytics or usage reporting

Sealed VBA orchestration — no user‑editable logic in the control layer

Compiled modules — core engine code is not exposed to workbook editing

This ensures that one tenant cannot alter the engine behavior for another, and that regulatory teams can rely on a stable, tamper‑resistant execution environment.

4. ML and Signals in a Multi‑Tenant Context

Machine‑learning signals, such as ML‑Enhanced Conviction, are computed and cached per tenant. Each tenant has its own:

ML cache file

signal export path

history consistency checks

feature engineering pipeline

Signals are never shared across tenants. This prevents one client’s history or labeling regime from influencing another client’s optimization outcomes.

5. Operational Benefits of Multi‑Tenant Design

The multi‑tenant architecture provides several practical advantages:

Regulatory separation — each client or desk can be isolated for compliance

Change control — configuration changes are scoped to a single tenant

Performance management — heavy workloads are contained within their own environment

Auditability — each tenant has its own deterministic history of runs and outputs

This makes ClearLedger suitable for firms that need strict separation between business units, strategies, or client books.

Conclusion

ClearLedger’s Multi‑Tenant Architecture combines local, in‑memory execution with hardened security and strict isolation of data, signals, and optimization workflows. Each tenant operates as a self‑contained analytical environment, with no shared cloud state and no cross‑tenant leakage. This design is a core strength of the ClearLedger engine for regulated portfolio teams.

← Back to Documentation Index