Network Topology 網路拓撲

Lightning Network Typical Metrics (Sample Data)

Network Scale:
• Public nodes: ~16,000+
• Public channels: ~75,000+
• Total capacity: ~5,000+ BTC
• Average channel capacity: ~0.07 BTC
• Median channel capacity: ~0.02 BTC

Connection Characteristics:
• Average channels per node: ~5-10
• Max connected node: 1000+ channels
• Average path length: ~3-4 hops
• Network diameter: ~6-8 hops

Note: Actual data changes over time, above are examples
Private channels not included in public statistics

Network Topology Patterns

1. Scale-Free Network
Characteristics:
• Few "super nodes" with many connections
• Most nodes have few connections

Pros:
• Short path length
• Efficient routing
• Resistant to random failures

Cons:
• Centralization risk
• Super nodes become single points of failure
• Privacy concentration

2. Hub-and-Spoke
      o
     /|\
    o-*-o   * = hub node
     /|\   o = edge nodes
      o

Real example: LSP and its users

3. Mesh Topology
o---o---o
| /| /|
| o | o |
|/ |/ |
o---o---o

Characteristics: Multiple paths between nodes
Pros: High redundancy, decentralized
Cons: Requires more locked funds

Network Centralization Metrics

Node Centrality Types:

Degree Centrality:
  Number of channels

Betweenness Centrality:
  How many shortest paths pass through node
  High betweenness = more payments route through

Closeness Centrality:
  Average distance to all other nodes
  High closeness = can quickly reach anywhere

Gini Coefficient:
Measures inequality in capacity/connection distribution
• 0 = perfect equality (all nodes same)
• 1 = perfect inequality (one node has everything)
• Lightning Network capacity Gini: ~0.8-0.9
  -> Highly concentrated, few nodes hold most capacity

Top N Node Control:
• Top 10 nodes control ~15% capacity
• Top 100 nodes control ~50% capacity
• Clear long-tail distribution

Topology Impact on Routing

Path Availability:
• High connectivity -> more route options
• Multiple paths -> MPP more likely to succeed
• Mesh structure -> single channel failure not fatal

Liquidity Distribution:
• Central nodes have high capacity -> large payments more likely
• Edge nodes have low capacity -> may need MPP
• Imbalanced channels -> need rebalancing

Latency and Reliability:
• Short path -> fewer hops -> lower latency
• Super nodes -> high availability -> more reliable routing
• But super node dependency -> larger failure impact

Fees:
• Competitive routes -> lower fees
• Chokepoints -> may charge higher fees
• Edge regions -> fewer choices, potentially higher fees

Strategies to Improve Network Structure

For Node Operators:

1. Selective Connections
   • Connect to nodes in different "regions"
   • Avoid connecting only to super nodes
   • Build triangle connections for redundancy

2. Capacity Allocation
   • Adjust channel size based on traffic
   • Maintain balanced inbound/outbound ratio
   • Allocate more capacity to frequent routes

3. Interconnection Coordination
   • Ring of Fire and similar community events
   • Establish channels with geographically diverse nodes
   • Support emerging nodes

Protocol Level Improvements:
• Channel Factories: multi-party shared funding tx
• Splicing: dynamically adjust channel capacity
• Trampoline: reduce end-node burden
• Better path discovery algorithms

Network Exploration Tools

Web Tools:
• 1ML (https://1ml.com) - node ranking, stats, search
• Amboss (https://amboss.space) - detailed node info, community
• LN+ (https://lightningnetwork.plus) - node connection coordination
• mempool.space/lightning - network overview

CLI Tools (LND):
# View network info
lncli getnetworkinfo

# View specific node
lncli getnodeinfo <pubkey>

# View node channels
lncli describegraph | jq '.nodes | length'

CLI Tools (CLN):
# Network statistics
lightning-cli listnodes | jq 'length'
lightning-cli listchannels | jq 'length'