# Node Architecture

Lattice operates as a two-tier network with distinct responsibilities:

| Tier                | Node Type      | Role                                                |
| ------------------- | -------------- | --------------------------------------------------- |
| **Control Plane**   | Orchestrators  | Coordinate work, validate proofs, manage reputation |
| **Execution Plane** | Resource Nodes | Provide actual infrastructure services              |

This separation enables specialization and economic efficiency. Orchestrators don't process application data. They coordinate and validate. Resource Nodes handle the actual work.

<figure><img src="https://content.gitbook.com/content/xy6Q4hi0tmcU8t5h3zJn/blobs/f3QBxCfKp1VwcmsTd7rD/image.png" alt=""><figcaption></figcaption></figure>

## How the Tiers Work Together

### Service Request Flow

1. **Application requests coordination** through Lattice
2. **Orchestrator receives request** and evaluates requirements
3. **Orchestrator selects Resource Nodes** based on reputation, location, capacity
4. **Resource Nodes provide services** (discovery, tunneling, relay)
5. **Orchestrator validates completion** through peer attestations
6. **Rewards distributed** based on verified contribution

### Validation Through Triangulation

When a Resource Node claims to have provided services (like tunneling traffic), Orchestrators don't take their word for it. Instead:

* Peers in the meshnet sign coordination events
* These signatures attest to Resource Node participation
* Orchestrators validate based on **peer perspectives**, not Resource Node telemetry

A tunnel cannot claim credit for traffic the peers didn't actually route through it.

## Network Services

Lattice provides services that individual meshnets cannot provide for themselves:

| Service           | Description                                                                     |
| ----------------- | ------------------------------------------------------------------------------- |
| **Discovery**     | Applications find available capacity without knowing node identities in advance |
| **NAT Traversal** | Hole punching and tunneling fallback for devices behind firewalls               |
| **Validation**    | Verify Proofs of Coordination against consensus rules                           |
| **Telemetry**     | Network-wide latency measurements for routing optimization                      |
| **Handshake**     | Key-value stores for exchanging connection details                              |
| **Proxy**         | Traffic relay when direct peer-to-peer fails (25-65% of traffic typically)      |

## Scale Design

Lattice is designed to scale to millions of nodes:

* **Geographic distribution** ensures low latency globally
* **Redundant Orchestrators** prevent single points of failure
* **Automatic failover** maintains availability when nodes fail
* **Reputation scoring** routes work to reliable infrastructure

{% hint style="warning" %}
The roles of Orchestration nodes and services of Resource nodes may change as demand is validated and optimized.
{% endhint %}