What a reversibility plan must actually contain
It is the most scrutinised piece of an outsourcing file, and the most often missing. A credible plan presupposes an architecture designed to be left.
Why the regulator starts there
In an outsourcing file, the reversibility plan describes how the institution retrieves its data and continues operating if the relationship ends. It is the piece that interests an examiner most, for a simple reason: it measures real dependency.
An institution that cannot leave its vendor has not outsourced a function, it has surrendered it. The distinction is structural, and it appears nowhere else in the contract.
The four elements that make a plan credible
FORMATS. An export in a proprietary format is not reversibility: it is moving house into a lorry you cannot open. The plan must name the formats, and they must be standard.
TIMESCALES. “On request” means nothing. A credible plan commits to a stated interval between request and delivery, and distinguishes the data export from the handover of documentation.
SCOPE. Operational data, but also decision traces, configuration, scorecards and policy versions. An institution that recovers its files without the policies that decided them cannot reconstruct a decision for an examiner.
ASSISTANCE. The plan must say who helps, for how long, and on what basis. Reversibility without support is theoretical reversibility.
What makes a plan credible upstream
A reversibility plan is not written: it is observed. It is credible if the architecture permits it, and declarative otherwise.
The properties that make it possible are technical and verifiable: standard components rather than proprietary services, externalised configuration rather than values baked into code, a database with no exclusive extension, documented AND TESTED export and import.
The important word is “tested”. A restoration procedure never executed is not a procedure, it is an intention.
A requirement that also protects the vendor
A vendor able to demonstrate reversibility gets into clients faster. The exit question, raised early and handled seriously, removes the heaviest objection a security committee has.
That is the useful paradox of this piece: the easier it is to leave, the easier it is to enter.
Sources
- WAMU Banking Commission — Circular no. 04-2017/CB/C on risk management.
- NEXCLARA Revision 1 — portability architecture, requirements E1 to E7.