Portable Data Architecture: Making Platform Choice Reversible in a Regulated Bank
Large banks have repeatedly tried to simplify fragmented data estates by consolidating onto a single warehouse, lake, cloud platform or marketplace. Each generation fixed the previous one's problems but moved the coupling elsewhere, and the premise that all data will end up on one target platform is increasingly unrealistic in a regulated, multi-entity, multi-jurisdiction group. This paper proposes a portable data architecture in which the data-product contract (ownership, legal entity, classification, purpose, residency, quality, service levels, retention and evidence) is defined independently of the platforms that serve it. Producers declare access modes in the contract and keep control of the implementation behind each. Consumers select from the modes they are entitled to use. A thin, logically coherent control plane determines which implementations are eligible and conformant, while existing governance and platform capabilities enforce policy. The paper sets this against data mesh, data fabric and marketplaces, relates it to BCBS 239, SS2/21, operational resilience and UK GDPR, and gives an illustrative real-options case for the value of reversibility. It is a proposal, not a result: it sets out a falsifiable experiment with two bindings, a defined acceptance test and an exit drill.
Authors
- Shaine Ismail
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-20
- DOI
- https://doi.org/10.5281/zenodo.24393941
- Primary Topic
- Information Technology Governance and Strategy
- Type
- article
- Field-Weighted Citation Impact
- 0.00