Cloud-Native Architecture

No servers. No racks. No compromises.

Sirona was built cloud-native from day one: unified infrastructure, a 99.95% uptime SLA, and zero required on-premise footprint. It's the foundation that makes AI-native radiology possible.

See Sirona in Action

How Sirona is Different

Cloud-native from day one, not retrofitted

Sirona was built cloud-native from the start, not migrated from an on-premise monolith. Every line of code, every data layer, every deployment pipeline assumes a cloud-first architecture. Most PACS platforms started as on-premise software and grafted cloud features on top. Sirona uses true microservices architecture: independent, deployable services that scale elastically. Reporter scales separately from viewer scales separately from AI orchestration. Continuous deployment ships new features continuously, without system downtime.

What cloud-native unlocks

Cloud-native isn't a deployment choice. It's a paradigm that determines what the platform can do, how fast it can evolve, and what AI is possible on top of it.

Microservices Architecture

Independent, deployable services scale elastically. Reporter, viewer, worklist, and AI layer each scale on their own demand profile.

RadOS

Continuous Deployment

New features ship continuously as invisible upgrades — no painful version migrations.

Learn more

Global CDN + Edge Streaming

Progressive image streaming from a global CDN delivers diagnostic-quality performance wherever your radiologists read — the same experience in the reading room and at home.

Edge streaming

99.95% Uptime SLA

Multi-AZ AWS redundancy, automatic recovery, continuous resilience testing. 99.99% measured historical uptime against a 99.95% contractual SLA.

Auto-scaling

Zero On-Prem Footprint

No PACS servers, no SAN arrays, no DR appliances. Just Chrome and a business-class internet connection.

Zero-footprint

Unified Data Model

All DICOM pixel data, all reports, all events share one data model — the prerequisite for agentic and multimodal AI.

Learn more

The cloud-native paradigm, in practice

Operations

No infrastructure to run on site

No servers to buy, no storage to provision, no patches to deploy, no DR plan to run. What is left on site is Chrome and, at most sites, an edge router.

No servers, SAN, or DR appliances

No patch cycles, no scheduled-downtime windows

No servers, storage, or workstation images to maintain

Elastic AWS capacity instead of hardware sized for peak

Global Reach

The same read, wherever your radiologists sit

Browser-based delivery from a global CDN means you read the same way regardless of geography — the opposite of legacy PACS, where distance from the server degrades the experience.

Progressive image streaming from a global CDN

Multi-AZ high availability on AWS

Reads over a standard business-class connection

Same experience wherever your radiologists sit

AI Foundation

If it's not cloud-native, it isn't AI-native

Agentic and multimodal AI require a unified data model, a shared event stream, and elastic compute — exactly what cloud-native delivers and what on-premise stacks struggle to retrofit. Cloud-native is the architectural entry condition for the next decade of radiology AI.

Unified data model across images and text

Shared workflow event stream for AI agents

Elastic compute for foundation models

AI-native is a cloud-native consequence

Resilience

99.95% uptime SLA — without maintenance windows

Multi-AZ redundancy, automatic in-region recovery, and continuous deployment mean Sirona stays up through infrastructure outages and vendor incidents. Multi-region deployment (US-East, UK, Australia, South Africa) arrives in the first half of 2027. The platform is designed to withstand the failures legacy PACS plans for with downtime.

99.99% historical uptime, 99.95% SLA

Multi-AZ AWS redundancy today

Multi-region by H1 2027 (US-East, UK, AU, SA)

No planned maintenance windows

99.95%

uptime SLA — 99.99% historical

PACS servers, SANs, or DR appliances on site

architecture, cloud-native from day one

What cloud-native looks like in practice

How cloud-native improves inter-practice collaboration

A fully managed path to integration

FAQs

What does 'cloud-native' actually mean here?

Built for the cloud from day one — microservices, continuous deployment, multi-AZ redundancy, and a unified data model. Not an on-prem system with a VPN bolted on.

Where does the data live?

Today, data is hosted in AWS with multi-AZ redundancy and automatic in-region replication. Additional regions and data-residency options are on our roadmap for markets that require them.

Can we read from home?

Yes. Radiologists read from home every day. Sirona runs in Google Chrome with the Sirona RadOS Manager extension over a business-class internet connection (we recommend 50 Mbps or better), with no VPN client or special hardware on the reading side.

What happens if an AWS data center fails?

Multi-AZ redundancy keeps you running through the loss of an availability zone — physically separated data centers within the region. Cross-region failover is on our resilience roadmap.

Do we still need any on-prem footprint?

No PACS servers, no SAN, no DR appliances. Readers need only Chrome, with no VPN client. Sites send images and HL7 over an encrypted tunnel, which most sites terminate on a small edge router.

What is Sirona's HIPAA and SOC 2 posture?

Sirona is HIPAA-compliant and runs an active SOC 2 Type II program; the attestation is available under NDA.