No masters. No slaves. No single point of failure. Every node is an equal peer that discovers routes dynamically and heals the network automatically.
A SaaS company runs data processing clusters across 5 regions (US, EU, Asia, AWS, on-prem). Networks are unreliable, nodes fail regularly, and adding/removing infrastructure is frequent. Traditional hierarchical systems fail.
No Kubernetes, no Consul, no etcd. Just nodes finding each other.
Time 0:00 - cluster-node-3 becomes unresponsive (network partition or crash)
Time 0:05 - Other nodes notice heartbeat missing
Time 0:10 - Routes recalculated, traffic rerouted
Time 0:15 - All active nodes have updated mesh view
Result: Zero manual intervention. Orchestrations continue across remaining nodes.
This is critical for production: you don't want oncall getting paged because one node is slow.
// HEALTHY state
[Mesh Online] - 5/5 nodes active
node-1: HEALTHY (route: 0 hops)
node-2: HEALTHY (route: 1 hop via node-1)
node-3: HEALTHY (route: 1 hop via node-1)
node-4: HEALTHY (route: 2 hops via node-1, node-2)
node-5: HEALTHY (route: 1 hop via node-4)
// DEGRADED state (node-3 fails)
[Mesh Degraded] - 4/5 nodes active
node-1: HEALTHY (route: 0 hops)
node-2: HEALTHY (route: 1 hop via node-1)
node-3: UNREACHABLE ⚠️
node-4: HEALTHY (route: 2 hops via node-1, node-2)
node-5: HEALTHY (route: 1 hop via node-4)
// RECOVERED state (node-3 rejoin)
[Mesh Online] - 5/5 nodes active
Routing tables updated automatically
No replay required
No manual reconciliationEach node acts as both a client and proxy:
Unlike Ansible Tower (centralized control node model), Dimensigon's peer-to-peer approach means your infrastructure is resilient to node failures.
// Ansible Tower Model (Centralized)
Control Node (Tower) → ssh → Node-1
↓ ssh → Node-2
└─ Single point of failure!
// Dimensigon Model (Decentralized)
Node-1 ←→ Node-2 ←→ Node-3
↓ ↓ ↓
Node-4 ←→ Node-5 ←→ Node-6
// Execute on Node-6 from Node-1
$ dshell orch run deploy-app --target=node-6
// Mesh automatically routes:
// Node-1 → Node-2 → Node-3 → Node-6
// (or any other available path)
// No central control node needed!# Bootstrap cluster (can be any node)
$ dimensigon new production
# Gen token on on-prem node
$ dimensigon token --expire 3600
# Join AWS nodes from CLI
$ dimensigon join on-prem-node.internal:20194 <TOKEN>
# Join Azure nodes
$ dimensigon join on-prem-node.internal:20194 <TOKEN>
# Join GCP nodes
$ dimensigon join on-prem-node.internal:20194 <TOKEN>
# View the mesh
$ dimensigon status
✓ MESH ACTIVE
Nodes: 8
Topology: Fully Connected
Redundancy: 3+ hops
Latency: on-prem→AWS 45ms, on-prem→Azure 65ms, on-prem→GCP 120msThe mesh doesn't care where nodes are. It just connects them efficiently.
ANSIBLE TOWER (Centralized Control)
┌─────────────────┐
│ Control Node │ ← Single point of failure
│ (Tower) │ ← Must be highly available
└────────┬────────┘
│ SSH to each node
┌────┼────┬─────┐
▼ ▼ ▼ ▼
Node1 Node2 Node3 Node4
Issues with Silos:
• Separate Tower instances for each silo
• Complex inter-silo communication
• Manual proxy configuration
• No automatic failover between silos
DIMENSIGON (Decentralized Mesh)
Region-1 Region-2
(AWS) (Azure)
┌─────┐ ┌─────┐
│ N-1 │◄──┐ │ N-5 │
│ N-2 │ ├─►│ N-6 │
│ N-3 │◄──┤ │ N-7 │
└─────┘ │ └─────┘
└─ Gateway Nodes
(Transparent routing)
Silo Interconnection Benefits:
✓ All nodes equal (no control node)
✓ Automatic gateway discovery
✓ Self-healing on node failure
✓ No manual firewall rules
✓ Resilient to region failuresMulti-Silo Problem: Traditional setups with Ansible Tower create isolated silos requiring manual bridge configuration.
Dimensigon interconnects silos elegantly. No control nodes. No manual bridges. Just mesh.
Mesh automatically elects gateway nodes to bridge silos. If a gateway fails, another is elected instantly.
// Node health check (auto-run)
"gateway_election": {
"health_check_interval": 5,
"failover_timeout": 10,
"elected_gateway": "node-5"
}Route messages through optimal paths considering latency, bandwidth, and reliability metrics.
// Auto-select best route
"aws-us" → "azure-eu": {
"direct": 85ms,
"via_gcp": 65ms,
"chosen": "via_gcp"
}All inter-node communication encrypted with mutual TLS. Encryption handled transparently at mesh layer.
"mesh_tls": {
"enabled": "true",
"cert_rotation": "90d",
"cipher": "TLS_1_3"
}Distribute orchestrations across nodes based on CPU, memory, and network capacity in real-time.
"selector": "region:aws"
AND "cpu<50%",
"matched_nodes": [
"aws-1", "aws-3"
]Network failures are handled automatically. Nodes rejoin when they recover.
Routes are learned dynamically. No manual IP/DNS management.
When a node fails, alternate routes activate in milliseconds.
Works on any cloud, on-prem, or hybrid setup without modification.