Cloud migration
From legacy Linux VMs to a modern, supportable public-cloud target architecture. Azure and AWS are shown as candidate options — neither is yet committed.
Legacy Linux VMs hosted on-premises with limited horizontal scaling.
Manual deployment processes and inconsistent observability across hubs.
Backup procedures in place but recovery time objectives are not regularly tested.
Containerised processing services on a managed cloud platform.
Centralised observability, SLO-driven operations and automated CI/CD.
Candidate platforms:Microsoft AzureAWS
Data ingestion
Managed event ingestion with schema registry, replayable streams.
Processing services
Containerised microservices, autoscaled per hub workload.
Storage
Object storage with tiering and lifecycle, plus relational metadata store.
Monitoring
Metrics, dashboards and SLO tracking per datahub and per pipeline stage.
Logging
Centralised structured logs with retention aligned to audit requirements.
Alerting
Severity-based routing to on-call rotations with escalation policies.
Security / IAM
Least-privilege roles, federated identity, secrets management, key rotation.
Backup & recovery
Cross-region snapshots, tested restore procedures, documented RPO/RTO.
- 1
Discovery
Service mapping, dependency capture, risk register.
- 2
Domain Onboarding
Engagement with EUMETNET stakeholders and member feeds.
- 3
Target Architecture
Reference architecture sign-off (Azure / AWS candidate options).
- 4
Platform Foundation
Landing zone, IAM, observability, CI/CD baselines.
- 5
E-GVap Redevelopment
Highest legacy footprint — earliest redevelopment target.
- 6
E-Profile Redevelopment
Containerised processing services rolled out.
- 7
E-ADAS Redevelopment
Phased migration with backwards-compatible interfaces.
- 8
Parallel Run
Side-by-side operation with reconciliation reporting.
- 9
Go-Live & Support
Cutover, hypercare, transition to steady-state operations.