Policy-as-Code and Fine-Grained Data Access Governance in Multi-Tenant Financial Platforms

Authors

  • Olabode Michael Soneye Ontario Ministry of Children, Community and Social Services (via Adisoft Inc.), Toronto, ON, Canada Author
  • Emmanuel Chukwuma Okeke Arcesium, New York, NY, USA Author
  • Ijeoma Mbonu Cloud and Cybersecurity Engineer, Dallas, TX, USA Author
  • Oluranti Ogundapo GTA, Nokia, Verizon Wireless Project, Canada Author
  • Funmilayo Ashore University of Georgia, Terry College of Business, Athens, GA, USA Author
  • Kofi Okyere Mensah 6CIBC Bank USA, USA Author
  • Ifeanyichukwu Okwesa Barclays Author

Keywords:

policy-as-code, authorization, ABAC, ReBAC, multi-tenancy, Open Policy Agent, Cedar, Zanzibar, row-level security, data governance, financial services, DORA, PCI DSS

Abstract

Multi-tenant financial platforms, including banking-as-a-service providers, payment processors, wealth and brokerage platforms, and core-banking SaaS, must enforce authorization decisions that are simultaneously fine-grained, tenant-isolated, auditable, and fast. To date, three converging forces reshaped how these platforms approach the problem: the maturation of declarative authorization languages and engines (Open Policy Agent and Rego, Cedar, and Zanzibar-style relationship engines such as OpenFGA); the extension of governance controls into the data plane through row-level security, column masking, and unified catalogs; and a regulatory calendar that placed explicit, testable obligations on access control, most notably the retirement of PCI DSS v3.2.1 and the run-up to the EU Digital Operational Resilience Act (DORA). This paper synthesizes the current state of practice into a taxonomy of five architectural models for policy enforcement: centralized decision service, distributed sidecar, embedded library, data-layer native enforcement, and relationship-graph authorization. It then analyzes them against a common set of security properties. We argue that no single model is dominant; the defensible design for regulated multi-tenancy is a layered hybrid in which tenant isolation is enforced as a non-negotiable invariant at the lowest layer, while business-level entitlements are expressed as versioned, analyzable policy higher in the stack. We identify the principal trade-offs (consistency versus latency, expressiveness versus analyzability, centralization versus blast radius, and fail-closed safety versus availability) and present a reference architecture, a threat model, and a verification regime suited to financial supervision.

Downloads

Published

2024-11-28

Similar Articles

11-20 of 35

You may also start an advanced similarity search for this article.

Most read articles by the same author(s)