Product  /  Platform & Runtime

A cloud-native runtime, not a refit of yesterday’s ESB.

Art2link was built from the ground up for Kubernetes — not packaged for it after the fact. Stateless workers, declarative deployments, per-pipeline scaling, and the operational ergonomics modern platform teams expect.

99.95%
Managed SLA
14
Regions
3
Deployment models
k8s
Native runtime
Deployment

Three ways to run Art2link. Same product, different boundary.

Pick the deployment that matches your security posture. The control plane, data plane, and customer data live in different places depending on which model you choose — toggle below to see each.

Managed

Art2link runs the runtime. You configure pipelines via console or API. Lowest operational burden, highest velocity. Most customers start here.

  • 99.95% SLA in 14 regions
  • Patching, scaling, DR all handled
  • Azure Marketplace one-click install
  • Customer-managed keys (CMK) supported

Hybrid

Control plane in Art2link Cloud. Data plane — including all message payloads — runs inside your VPC and never crosses the boundary. The right choice for regulated workloads.

  • HL7, X12, PII never leave your network
  • You hold the encryption keys
  • Air-gap-friendly config sync
  • HIPAA / GLBA / PCI workloads

Self-hosted

The entire stack runs in your infrastructure — air-gapped, on-prem, sovereign cloud. Helm chart, offline activation, no inbound call-home.

  • Bring-your-own Kubernetes
  • Offline license activation
  • Government / defense / sovereign-cloud
  • Long-term support releases (LTS)
Runtime

Stateless workers on Kubernetes.

Every pipeline runs as a pool of stateless workers. Crashes are non-events — pods restart, the message broker holds state, no in-flight work is lost.

Scaling

Per-pipeline autoscale.

Each pipeline scales independently. The high-throughput EDI lane doesn’t starve the low-latency HL7 lane. KEDA-driven, queue-depth-aware.

Storage

Audit-first, no message hoarding.

Messages persist only long enough to deliver — configurable retention per pipeline. Audit metadata is long-lived and immutable. Right-to-be-forgotten compliant.

Multi-region

Active-active across geographies.

Pipelines can span regions for disaster tolerance. Cross-region replication is opt-in per pipeline. Failover is config, not surgery.

Networking

Private by default.

VNet integration, Private Link, AWS PrivateLink, customer-managed IP allowlists. No public ingress required for any production pipeline.

Versioning

Pipelines as code, deployments as artifacts.

Every pipeline change is a signed deployment artifact. Roll forward, roll back, A/B route — same primitives engineers expect for code.

99.95%
Uptime SLA
< 50ms
p99 routing latency
14
Regions
24/7
On-call coverage

Talk to a platform architect.

Show us the topology you’re running today — we’ll sketch the Art2link version on a 30-minute call. No slides.

Book architecture review  →