Pilot Programme Proposal

Phase 0: A 24-36 month proof-of-concept with €200-500M budget, specific success criteria, and go/no-go gates before committing to full programme.

PAC/NAO Challenge

"You want billions committed before any proof this approach works? Where's the pilot? Where's the gateway review?"

This document addresses this by proposing a structured pilot phase with independent validation before major investment decisions.

Strategic Framing: Insurance, Not Imminent Response

This initiative is strategic risk reduction, not a response to an imminent threat. The probability of US service termination to allied governments is low under normal conditions. However, the impact of such an event would be catastrophic. This programme is insurance: reducing dependency and leverage while maintaining partnership. The pilot validates whether this insurance is technically achievable and cost-effective before major commitment.


1. Pilot Programme Overview

Aspect Specification
Duration 24-36 months (with 6-month extension option if needed)
Budget €200-500M across all participating jurisdictions
Scope 2-3 non-critical government services per jurisdiction
Governance Lightweight structure reporting to national CIOs
Decision point Full Cabinet/Treasury sign-off required at each gate
Exit criteria Clear conditions for terminating pilot if unsuccessful

Pilot Budget Breakdown

€80-120M Infrastructure (2-4 DC nodes)
€60-100M Platform Development
€40-80M Migration & Testing
€20-50M Programme Management

UK share: Approximately £80-150M (based on 25% contribution weight)

2. Gateway Review Structure

The pilot follows IPA Gateway Review principles with five formal decision points.

G0

Strategic Assessment

Month 0
Programme approval

G1

Business Case

Month 6
Detailed design

G2

Delivery Strategy

Month 12
Procurement complete

G3

Investment Decision

Month 24
Pilot results

G4

Full Programme

Month 30-36
Go/No-Go

Gateway Decision Matrix

Gate Key Questions Decision Authority Possible Outcomes
G0 Is there a valid strategic case for sovereign cloud? Cabinet/National Cabinet Approve pilot / Reject
G1 Is the technical approach sound? Are costs realistic?
HARD PRECONDITION: Has JIC/NCSC threat assessment validated the strategic rationale?
Senior Officials + IPA review Proceed / Revise / Stop
Cannot proceed without intelligence validation
G2 Are vendors capable? Are contracts value for money? Procurement authority + IPA Award contracts / Re-tender / Stop
G3 Has the pilot demonstrated capability? Are services functional? Technical Board + IPA Proceed to full programme / Extend pilot / Stop
G4 Is full programme investment justified? Are risks acceptable? Cabinet/Treasury Approve full programme / Phased approach / Terminate

3. Pilot Service Selection

Pilot services are selected based on: (1) non-critical nature (failure doesn't affect citizens), (2) representative workload patterns, (3) known technology requirements, (4) willingness of owning department to participate.

UK Proposed Pilot Services

1. GOV.UK Static Content (Tier: Low risk)

Description: Static informational pages from GOV.UK (not transactional services)
Technology: Object storage, CDN, basic compute
Success criteria: 99.9% uptime, <100ms latency, successful content sync

2. Internal HR Portal (Tier: Low risk)

Description: Civil Service HR self-service for leave requests, directory
Technology: Web app, relational database, authentication (Keycloak)
Success criteria: Feature parity, successful login integration, data integrity

3. Open Data Portal (Tier: Medium risk)

Description: data.gov.uk or subset, demonstrating data services
Technology: Object storage, APIs, search (PostgreSQL), analytics
Success criteria: API compatibility, query performance, data freshness

EU Proposed Pilot Services

1. Europa.eu Information Pages

Static content from European Commission websites

2. EU Open Data Portal

data.europa.eu demonstrating cross-border data access

Canada Proposed Pilot Services

1. Canada.ca Static Pages

Informational content from federal portal

2. Open Government Portal

open.canada.ca data catalogue

Australia Proposed Pilot Services

1. data.gov.au

Australian open data portal

2. Non-production myGov sandbox

Testing environment (not production)

4. Success Criteria and Metrics

Quantified Success Criteria for Gate 3

All criteria must be met for "Proceed" recommendation. Partial achievement triggers "Extend pilot" recommendation. Major failure triggers "Stop" recommendation.

Technical Success Criteria

Metric Target Measurement Method
Platform availability ≥99.9% uptime (8.76 hours max downtime/year) Automated monitoring, monthly reports
Service latency P95 <200ms for web applications Application performance monitoring
Migration success 100% of pilot services migrated without data loss Data integrity verification, user acceptance testing
Interoperability Cross-jurisdiction API calls successful 99%+ Automated integration testing
Security posture Zero critical/high vulnerabilities unpatched >30 days Continuous vulnerability scanning, pen testing
Recovery capability RTO <4 hours, RPO <1 hour demonstrated Quarterly DR exercises

Operational Success Criteria

Metric Target Measurement Method
Cost variance Within 20% of budget Monthly financial reporting
Schedule adherence Key milestones within 3 months of plan Project tracking, IPA review
Vendor performance All SLAs met >95% of time Contract management, SLA reporting
Team capability Core platform team trained and retained Skills assessment, retention metrics

Coordination Success Criteria

Metric Target Measurement Method
Governance effectiveness No deadlocks exceeding 30 days Decision log, escalation records
Standard adoption All jurisdictions using common APIs and formats Interoperability testing
Knowledge sharing Documented lessons learned, shared tooling Repository activity, cross-party contributions

5. Go/No-Go Decision Framework

At Gate 4, the following decision framework applies:

Outcome Criteria Next Steps
GO: Full Programme All technical criteria met
Cost within 20% of budget
No major governance failures
All jurisdictions confirm continued participation
Proceed to full business case
Submit for Cabinet/Treasury approval
Begin Phase 1 procurement
CONDITIONAL: Phased Approach Most technical criteria met with issues
Cost 20-40% over budget
Some governance challenges resolved
3+ jurisdictions confirm participation
Proceed with reduced scope
Address specific issues
Additional 12-month phase before full programme
EXTEND: Additional Pilot Technical criteria partially met
Identified issues are fixable
Political will remains
12-month pilot extension
Additional €50-100M budget
Revised success criteria
NO-GO: Terminate Multiple technical failures
Cost >50% over budget
Unresolvable governance breakdown
Any jurisdiction withdraws
Orderly wind-down (12 months)
Data returned to original systems
Lessons learned report
Consider alternative approaches

Independent Validation

Go/No-Go decision must be supported by:

  • IPA Gateway Review: Independent review by Infrastructure and Projects Authority
  • NAO Value for Money: NAO assessment of pilot costs vs. outcomes
  • Technical Expert Panel: Independent technical review (3 external experts)
  • User feedback: Civil servants using pilot services surveyed

6. Pilot-Specific Risks and Mitigations

Risk Probability Impact Mitigation
CloudStack doesn't meet government security requirements Medium High Security assessment at Gate 1; fallback to OCI Dedicated as comparison
Open-source FaaS (OpenFaaS/Knative) performance inadequate Medium Medium Benchmark against AWS Lambda; accept higher latency or develop optimisation
Multinational coordination proves unworkable Low High Lightweight governance for pilot; can fall back to national-only approach
Vendor market doesn't respond to tenders Low High Market engagement before tender; consider in-house capability
Political change withdraws support Medium High Cross-party briefings; frame as national security not political project
Pilot succeeds but full programme fails due to scale Medium High Pilot must include scalability testing; stress tests at 10x load

Summary

Aspect Value
Duration 24-36 months
Budget (total) €200-500M
UK share ~£80-150M
Services piloted 2-3 per jurisdiction (8-12 total)
Gateway reviews 5 formal gates with IPA involvement
Go/No-Go decision Month 30-36 with Cabinet/Treasury approval
Exit option Terminate at any gate with orderly wind-down

Document Status

This pilot programme proposal provides the "prove before committing" structure that addresses PAC, NAO, and Treasury concerns about large-scale IT investment.

Version: 1.0 | Last updated: January 2026

Back to main documentation