Deployment Flexibility
From a single node at a forward location to multi-site constellations spanning continents—Starlight scales to your operational requirements.
Traditional orchestration platforms depend on central control planes. When that control plane is unreachable, operations stop. Starlight inverts this model.
Every Starlight node operates as a fully autonomous unit capable of running workloads, making scheduling decisions, and maintaining security posture—regardless of network connectivity to other nodes or centralized management.
This isn't a fallback mode. It's how Starlight operates by design. Connectivity enables coordination; it doesn't enable operation.
Single-Node Edge
Full platform capability on a single server at forward locations
Multi-Node Clusters
High availability within a single location
Distributed Constellations
Coordinated operations across multiple sites and regions
Data Center Density
Scales to thousands of VMs when density is required
Distributed, Worker-First Design
Architecture Principles
No Single Point of Failure
Every node operates autonomously. Loss of any node—or any connection between nodes—doesn't stop operations elsewhere.
Eventually Consistent
State coordination happens when possible, not when required. Nodes reconcile state when connectivity returns without operator intervention.
Open Standards
Built on CNCF-aligned technologies. Standard interfaces mean standard skills apply and standard tools integrate.
Local Authority
Decisions are made at the edge, not in a distant data center. Latency-sensitive operations don't wait for round-trips.
Resilient by Default
Designed for contested and degraded environments. Network partitions, node failures, and adversary actions are expected conditions.
Immutable Foundation
The operating system and platform components are immutable. No configuration drift, no accumulated state, no mystery about what's running.
Learn About Security
Discover how Starlight protects workloads with confidential computing and Zero Trust.