Execution nodes
Run approved agent workloads inside operator-defined trust boundaries.
NODE NETWORK
Distributed resilience without surrendering operator sovereignty.
The planned SANDA node architecture is intended to distribute execution, monitoring, containment, and recovery across approved infrastructure. The design goal is continuity without creating another opaque centralized dependency.
PLANNED NODE ROLES
Different responsibilities. Explicit authority.
A resilient network is not a pile of identical servers. Each node type should have bounded permissions and a defined failure role.
Sentinel nodes
Observe policy state, health, anomalies, and risk signals without taking unrestricted control.
Recovery nodes
Support safe failover, restoration, and continuity when primary infrastructure becomes unavailable.
Research nodes
Participate in isolated simulation, validation, benchmarking, and adversarial testing environments.
NETWORK CLAIM GUARDRAILS
Credibility depends on saying what is not yet proven.
These constraints are part of the product strategy, not footnotes added after launch.
- No public network-scale claim is presented as achieved until independently verified.
- Node roles, permissions, and trust boundaries must be explicit before deployment.
- Sensitive workloads should never be distributed merely for marketing optics.
- Operator authority, revocation, and containment remain first-class design requirements.
Design resilience around the workflow—not a vanity node count.
A node pilot begins with roles, permissions, revocation, containment, and failure recovery.