P2P Network Module¶

The P2P Network module contains approximately 27 models that analyze the peer-to-peer network infrastructure of Gnosis Chain. It covers node discovery protocols (discv4 and discv5), network topology, geographic distribution of nodes, client version tracking, ISP diversity, and connectivity metrics.

Data Sources¶

P2P data is sourced from two ClickHouse databases:

• nebula -- P2P network crawl results and visit records from the Nebula crawler (discv5)
• nebula_discv4 -- Discovery v4 protocol-specific data (variant schema)

The Nebula crawler periodically discovers and visits nodes on the Gnosis Chain network, recording their capabilities, geographic location (via IP geolocation), client identifiers, ISP information, and protocol support.

Key Concepts¶

Nebula Crawler

An open-source network measurement tool that systematically discovers and visits P2P nodes. Each crawl session attempts to connect to all known nodes and records reachability, latency, client agent string, and supported protocols. Crawl frequency is approximately every 30 minutes.

discv4 vs discv5

Two generations of the Ethereum node discovery protocol.

• discv4 uses a Kademlia-based DHT for peer discovery. It is the legacy protocol still used by older clients and is simpler but less feature-rich.
• discv5 is the newer protocol supporting topic advertisement, improved NAT traversal, and better ENR (Ethereum Node Record) handling. Most modern clients support both protocols.

Agent String

The self-reported client identifier sent during the P2P handshake (e.g., Nethermind/v1.25.0/linux-x64). Agent strings are parsed to extract structured fields: client name, version, platform, and runtime. This enables tracking client diversity and upgrade adoption rates.

Reachable vs. Total Nodes

Total nodes include all peers discovered in the DHT. Reachable nodes are those that responded to a connection attempt during the crawl. The ratio of reachable to total nodes is a health indicator for the network.

Model Categories¶

Clients

Discovery

Discv4

Discv5

Neighbors

Topology

Visits

Network Health Metrics¶

The P2P module provides several network health indicators:

Key Models Reference¶

Query Examples¶

Retrieve daily network size for the past 30 days:

SELECT dt, total_nodes, reachable_nodes
FROM dbt.api_p2p_network_size_daily
WHERE dt >= today() - 30
ORDER BY dt

Check client distribution:

SELECT dt, client_name, node_count
FROM dbt.api_p2p_nodes_by_client_daily
WHERE dt = today() - 1
ORDER BY node_count DESC

Analyze geographic spread:

SELECT country, node_count,
       round(node_count / sum(node_count) OVER () * 100, 2) AS pct_of_total
FROM dbt.api_p2p_nodes_by_country_daily
WHERE dt = today() - 1
ORDER BY node_count DESC
LIMIT 20

Track version adoption after a client release:

SELECT dt, client_version, node_count
FROM dbt.int_p2p_nodes_by_client_version_daily
WHERE client_name = 'Nethermind'
  AND dt >= today() - 14
ORDER BY dt, node_count DESC

Related Modules¶

• Consensus -- Validator data correlates with consensus client distribution
• ESG -- Geographic node distribution feeds energy consumption estimates
• ProbeLab -- Additional network monitoring with cloud provider classification