Skip to content
Control Plane

Architecture

The Control Plane platform suite

Five components that fit together — managed Kubernetes for any compute, a serverless layer that hides the K8s complexity, a bring-your-own variant for sovereign infrastructure, a pooled GPU marketplace, and the white-glove services that get you there.

Sign Up for FreeTalk to an Engineer

CPLN-MK8s · delivered from the cloud

Control PlaneManaged Kubernetes · one cluster, unified API
UpgradesAutoscalingSpot cost optimizationAdd-onsCNCF Certified

One managed cluster across your compute

Bare MetalOn-prem VMsAWSGCPAzureHetzner

01 · Managed Kubernetes, anywhere you have compute

CPLN-MK8s

MK8s delivers a Kubernetes cluster from the cloud to manage bare metal hardware, on-premises virtual machines, or any cloud provider's infrastructure as part of a single managed cluster. Built for environments that lack native K8s facilities — or that need a unified API for deployment and monitoring across mixed compute.

  • One managed cluster across bare metal, on-prem VMs, and any cloud
  • Unified API for deploying, upgrading, and continuously monitoring
  • Add-ons: dashboards, workload identity, logs & metrics, fine-grained authorization
  • Autoscaling, spot-instance cost optimization, and seamless upgrades
  • CNCF Certified Kubernetes conformance

CPLN-Serverless · five symmetric interfaces

MCPUICLIIaCAPI
Control PlaneYour Virtual Cloud · any region, any cloud
Geo-routingAuto-scalingCost optimizationSecretsService discoveryThreat detectionImage scanningService mesh · mTLSCloud Wormhole®Universal Cloud Identity™

$ cpln workload create --name api-gateway --image my-org/api:latest --type serverless -o json

{ "kind": "workload", "name": "api-gateway", ... }

02 · Kubernetes without the complexity of Kubernetes

CPLN-Serverless

Run workloads across any number of regions and clouds through five symmetric interfaces — UI, API, CLI, Terraform/Tofu, and Pulumi. Geo-routing, auto-scaling, cost optimization, secrets, service discovery, threat detection, and container scanning are built in. Cloud Wormhole® provides cross-network tunneling and Universal Cloud Identity™ conveys identity across clouds, so workloads stay portable — between clouds, or from on-premises to cloud and back. An integrated service mesh provides mutual TLS via Envoy and Istio.

  • Geo-routing across regions and clouds
  • Auto-scaling, scale-to-zero, and continuous cost optimization
  • Secrets management & service discovery built in
  • Threat detection & container image scanning
  • Mutual-TLS service mesh via Envoy and Istio
  • Cloud Wormhole® for cross-network tunneling
  • Universal Cloud Identity™ for cross-cloud identity
  • Workload portability between clouds and on-prem

CPLN-BYOK · same control plane

Control PlaneEvery CPLN-Serverless capability, unchanged
WorkloadsNetworkingIdentityObservabilityPolicy

Your infrastructure · registers as a location

Your KubernetesBare MetalVirtualizedEdge
SovereigntyRegulated workloadsYour accounts

03 · Bring-your-own-Kubernetes — same platform, your compute

CPLN-BYOK

Every CPLN-Serverless capability, but executed on customer-controlled compute infrastructure instead of Control Plane facilities. Run on bare metal, virtualized hardware, or provider-owned resources — keep the operational model, choose where the workloads land.

  • Same control plane, your underlying compute
  • Your cluster registers as a location — workloads roll out to it like any region
  • Bare metal, virtualized, or provider-owned compute
  • Inherits every CPLN-Serverless feature
  • Useful for sovereignty and regulated workloads

CPLN-GPU-Ocean℠ · one supply, many providers

Control PlaneGPU Ocean℠ · pooled procurement & management
ProcurementManagementDisposal

Excess-capacity providers · reserved instances & bare metal

GPUB200GPUH200GPUH100GPUA100GPUL40STPUTPU v5e+ more

04 · Pool GPU and TPU capacity from many providers

CPLN-GPU-Ocean℠

GPU Ocean℠ is a cloud computing management solution that pools GPUs, TPUs, and similar advanced computing nodes drawn from excess-capacity providers for resale. It simplifies procurement and management of GPU workloads across diverse, fragmented infrastructure.

  • Pool GPUs, TPUs, and accelerator capacity
  • Excess capacity — reserved instances or bare metal — brokered as one supply
  • Procure, manage, and dispose of capacity wherever it runs
  • Managed through the same Control Plane interfaces

Additional services · white glove

White-glove migrationfrom legacy infrastructure to the platform
Containerization serviceslegacy apps onboarded without re-architecture
Premier supportpriority help from Control Plane engineers
Control PlaneYour workloads, running on the platform

05 · Migration, premier support, containerization

Additional Services

Beyond the platform itself, Control Plane offers migration services, premier support, and containerization services — helping customers modernize their cloud and on-prem infrastructure and bring legacy applications onto the platform.

  • White-glove migration from legacy infrastructure
  • Containerization services for non-containerized apps
  • Premier support from Control Plane engineers
  • Modernize cloud and on-prem infrastructure without re-architecture

Give your cloud superpowers

Sign up free. No cloud accounts required.

Start BuildingTalk to an Engineer
  • SOC 2 Type II
  • PCI DSS Level 1
  • HIPAA