I’ve previously talked about how migrating NHS Spine to the cloud will provide a secure, adaptable, and sustainable infrastructure for the health and care system. But this is a huge undertaking: it will be one of the largest cloud migrations that the NHS has ever undertaken.

My team’s challenge is how to migrate Spine’s 36 national services to the cloud. They are hosted on more than 360 servers across 2 UK datacentres and contain more than 1.2 petabytes of data. The migration must be achieved with zero down-time to ensure availability and continuity of services for all end users.

Back in April 2022, when the Spine Futures programme began, our initial plan was to migrate each Spine service to the cloud, transforming it where appropriate, to make best use of cloud technology and build independence between products and services. 

The Message Exchange for Social Care and Health (MESH) service was the first migration using this approach, and it went live in December last year. MESH was chosen as a relatively simple service, with minimal dependencies on other Spine systems. However, MESH proved significantly more challenging and complex than was first thought and continuing this approach for the remaining services would not be achievable within the 3-year programme.

In order to reduce the risk of multiple complex transition steps for each Spine product and accelerate the move to cloud, we plan on migrating all services together in one go and re-hosting them into AWS (Amazon Web Service) to run in EC2 instances. The live services cut over is due to start in October and conclude in early November 2023.  Once migrated, we will have opportunity to optimise Spine in the cloud and leverage the benefits of cloud native tools to help decouple and improve the services.

Early in the planning phase of the migration we agreed 2 simple, but fundamental principles that we need to follow for the migration to be successful:

1. Only make changes where absolutely necessary. We plan on making no functional changes to any Spine service during the migration. The only change will be the infrastructure that the code sits on. This reduces complexity and keeps us focused on our objective.  
2. Ensure the continuity of services for end users during migration. Our migration plan is designed to ensure continuity of service for all Spine products during the migration and all existing customer connections will be transferred over to AWS Cloud. There will be no change to the Spine IP addresses that are currently accessed.

Connectivity and regression testing

NHS Spine is a critical national service that health and care organisations depend on. To help ensure a successful migration to AWS Cloud, we have planned a 2-month period of connectivity and regression testing, starting on 1 July until 31 August 2023. This will be undertaken in the Spine integration environment.

We’re asking connecting organisations to take part and test their application’s connectivity and message responses to assure that there is no impact to the organisation or their customers. Organisations should validate that their applications can continue to connect to Spine and that inbound messages and receipt of outbound asynchronous messages are unaffected. Testing should include services like Personal Demographic Service (PDS), Summary Care Record (SCR), Electronic Prescription Service (EPS), the National Event Management Service (NEMS), Child Protection Information Service (CP-IS) and GP2GP.

So that we can monitor testing, we’d be grateful if those affected could email us at [email protected] to confirm your organisation's participation. I’d like to thank all those organisations that have already agreed to participate and help ensure that this migration is a success.

We have noticed that some suppliers have hardcoded their connection to Spine to a specific IP address.  This will cause issues during any site failovers on Spine.

As part of the cloud migration live service cut over, we’ll move IP addresses to the cloud.  Organisations that have hardcoded to a specific site will experience an outage.  We’re asking suppliers to use Spine’s Fully Qualified Domain Names (FQDN’s) and not specific IP addresses to avoid service disruption.  Not only will this ensure site failovers are handled dynamically, it will also ensure service continuity during the Spine transition, avoiding user impact and clinical risk.

We’ll be working with organisations over the coming weeks to help them prepare for a smooth transition to the cloud. More information will be shared soon.

You can visit our Spine Futures webpage to follow news and events.