The Provider

Providers are the backbone of the WorldLand network, contributing GPU compute resources and earning WL tokens in return.

What is a Provider?

A Provider is any entity that contributes GPU resources to the WorldLand network. Providers can be:

• Individual GPU owners
• Data centers
• Mining operations converting to compute
• Tech companies with idle GPU capacity
• Gaming studios with spare infrastructure

┌─────────────────────────────────────────────────────────────────┐
│                    Provider in WorldLand                        │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│   Provider                        WorldLand Network             │
│   ────────                        ─────────────────             │
│                                                                 │
│   ┌─────────────┐                 ┌─────────────────┐          │
│   │  GPU Node   │  ◀──────────▶   │  Broker   │          │
│   │ (Worker)    │   Register +    │  (Orchestrator) │          │
│   │             │   Heartbeat     │                 │          │
│   │ • RTX 4090  │                 └────────┬────────┘          │
│   │ • A100      │                          │                    │
│   │ • etc.      │                          ▼                    │
│   └─────────────┘                 ┌─────────────────┐          │
│         │                         │   Kubernetes    │          │
│         │ Join Cluster            │   Master Node   │          │
│         └────────────────────────▶│                 │          │
│                                   └─────────────────┘          │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Provider State

Providers go through a lifecycle of states:

┌──────────┐     ┌──────────┐     ┌──────────┐     ┌───────────┐
│ Pending  │ ──▶ │ Approved │ ──▶ │  Joined  │ ──▶ │ Available │
└──────────┘     └──────────┘     └──────────┘     └───────────┘
                                        │
                                        │ Heartbeat Timeout
                                        ▼
                                  ┌──────────┐
                                  │ Offline  │
                                  └──────────┘

Status	Description
Pending	Registration request submitted
Approved	Join token issued, awaiting cluster join
Joined	Successfully joined Kubernetes cluster
Available	Ready to accept GPU jobs
Offline	Heartbeat stale (> 2 minutes without heartbeat)

Provider Registration Flow

1. SDK Installation

Provider runs the WorldLand Provider SDK:

./worldland-provider-sdk install

2. Registration Message

SDK publishes a registration request to Redis:

{
  "provider_id": "provider-uuid",
  "wallet_addr": "0x1234...abcd",
  "spec": {
    "hostname": "gpu-server-01",
    "public_ip": "123.45.67.89",
    "total_gpus": 4,
    "gpus": [{ "name": "NVIDIA GeForce RTX 4090", "memory": 24576 }],
    "cpu_cores": 32,
    "total_memory_mb": 131072,
    "available_disk_gb": 500
  },
  "capacity": {
    "gpu_price_per_hour": 0.5
  }
}

3. Orchestrator Processing

The Broker's Orchestrator:

1. Validates the registration request
2. Generates a kubeadm join token
3. Stores provider state in memory and database
4. Sends join command to provider

4. Cluster Join

Provider executes the join command:

kubeadm join <master-ip>:6443 \
  --token <token> \
  --discovery-token-ca-cert-hash sha256:<hash>

5. Node Labels

Once joined, the node is labeled for GPU workloads:

labels:
  worldland.io/rental-type: gpu
  worldland.io/provider-id: provider-uuid
  worldland.io/gpu-model: RTX-4090

Provider Capacity

Each provider tracks resource capacity:

type ProviderCapacity struct {
    // GPU tracking
    AvailableGPUs    map[string]int  // Available by model
    InUseGPUs        map[string]int  // In use by rental jobs
    MiningGPUs       map[string]int  // Allocated to mining

    // CPU/Memory tracking
    TotalCPUCores      int
    AvailableCPUCores  int
    InUseCPUCores      int

    TotalMemoryMB      int64
    AvailableMemoryMB  int64
    InUseMemoryMB      int64

    // Pricing
    GPUPricePerHour    float64
}

Resource Allocation

When a customer creates a job:

1. AllocateResources() - Deducts GPU, CPU, Memory from available pool
2. Job runs - Container uses allocated resources
3. ReleaseResources() - Returns resources when job ends

Before Job:
  Available GPU: 4
  Available CPU: 32 cores
  Available Memory: 128 GB

Customer requests: 1 GPU, 8 CPU, 32GB RAM

After Allocation:
  Available GPU: 3 (-1)
  Available CPU: 24 cores (-8)
  Available Memory: 96 GB (-32)

  In-Use GPU: 1
  In-Use CPU: 8 cores
  In-Use Memory: 32 GB

Heartbeat System

Providers must send regular heartbeats:

type HeartbeatMessage struct {
    ProviderID  string
    NodeName    string
    Status      RegistrationStatus
    ActiveJobs  int
    Timestamp   time.Time
}

Monitoring

• Heartbeat interval: Every 30 seconds
• Stale threshold: 2 minutes without heartbeat → Offline
• Node labels updated: Active jobs, last heartbeat time

Mining Support

Providers can allocate GPUs for mining when not rented:

Total GPUs: 4
├── Available (for rental): 2
├── In-Use (rental jobs): 1
└── Mining: 1

Mining vs Rental Priority

• Rental jobs have higher priority
• Mining GPUs can be released for rental demand
• Mining resumes when GPUs become available

Provider Specification

Information collected from each provider:

Field	Description	Example
Hostname	System hostname	gpu-server-01
Public IP	External IP for SSH access	123.45.67.89
Total GPUs	Number of GPUs	4
GPU Info	Model, memory per GPU	RTX 4090, 24GB
CPU Cores	Total CPU cores	32
Total Memory	System RAM	128 GB
Available Disk	Free storage	500 GB

Next Steps

• The Broker - Central orchestration
• Network Participants - Roles in the network
• How to Provide - Setup guide