IPFS logo

Topology

Key metrics on network structure, geographic peer distribution, client diversity, and connection patterns across the IPFS network. All metrics in this section refer to IPFS Amino DHT Server nodes.

Geographic Distribution

Node IP address to country, city, and autonomous system (ASN) mapping, based on data sourced from https://ipregistry.co/.

Peers By City

Peers By City

This map shows the online peers per city in the last crawl of the latest day. The city and its coordinates come from the IP address mapping of ipregistry.co. Peers without a known city are not on the map. The text below the map shows how many peers are not on the map.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Loading ...

Geolocation Over Time

Geolocation Over Time

Shows the number of online peers per country in each crawl. The six countries with the most peers in the selected time range have their own series. Other holds the rest, including the peers without a known country. The country of a peer comes from the IP address mapping of ipregistry.co. Select an agent type to show only the peers of that type.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

DHT Server geolocation per agent version over time

Countries Distribution (24h)

Countries Distribution (24h)

Shows the online peers per country, as the average number per crawl in the 24 hours up to the latest crawl. The bars show the ten countries with the most peers by their two-letter country code. Hover over a bar to see the country name. other holds the rest, including the peers without a known country. The label on each bar is its share of all peers. The country of a peer comes from the IP address mapping of ipregistry.co. Select an agent type to show only the peers of that type.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Snapshot of active agents geolocation within the last 24hrs

Cities Over Time

Cities Over Time

Shows the number of online peers in the ten cities with the most peers over time. The other cities are in "Other". The city of a peer comes from the IP address mapping of ipregistry.co. Peers without a known city are not in this chart. Select an agent type to show only the peers of that type.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Online peers in the largest cities

Autonomous Systems Over Time

Autonomous Systems Over Time

Shows the number of online peers in the ten largest autonomous systems (ASNs) over time. The other ASNs are in "Other". The ASN of a peer comes from the IP address mapping of ipregistry.co. Peers without a known ASN are not in this chart. Select an agent type to show only the peers of that type.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Online peers in the largest autonomous systems (ASNs)

ASN Concentration

ASN Concentration

Shows how concentrated the online peers are in autonomous systems (ASNs). Only peers with a known ASN count. Select a view:

  • Largest ASNs: the share of the peers in the largest ASN and in the ten largest ASNs.
  • ASNs for 50%: the smallest number of ASNs that together host half of the peers. A low number means that a few network operators host much of the network.
  • HHI: the Herfindahl-Hirschman index of the ASN shares, from 0 (spread over many ASNs) to 10,000 (all peers in one ASN).

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

How many autonomous systems host the online peers

Agent Version Distributions

Network structure representation through client and agent distribution.

Client Diversity Trend

Client Diversity Trend

Shows the number of online peers per agent type (client software) in each crawl. The areas are stacked, so their top is the number of all online peers. The six agent types with the most peers in the selected time range have their own series. Other contains all other agent types.

The Unknown category represents nodes where the agent version could not be retrieved through the libp2p identify protocol. This situation can occur due to various reasons, such as the node terminating the connection before completing the full protocol negotiation.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Client participation over time

Active Agents Distribution (24h)

Active Agents Distribution (24h)

Shows the client software (agent type) of the online peers. One value is the average number of online peers per crawl in the 24 hours up to the latest crawl. The six agent types with the most peers are shown separately. Other contains all other agent types. The percentages are shares of all online peers.

Our Nebula crawler gets the agent version of a peer through the libp2p identify protocol. Unknown contains the peers where this did not work, for example because the peer closed the connection before the protocol negotiation completed.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Snapshot of active agents within the last 24hrs

Agent Versions Trend

Agent Versions Trend

Shows the number of online peers per version of one agent type (client software) in each crawl. The areas are stacked, so their top is the number of online peers of this agent type. Select the agent type in the menu. The menu lists the six agent types with the most peers in the selected time range, and the first one is the default. The six most common versions in the selected time range have their own series. Other contains all other versions and the peers where we cannot read a version from the agent version.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Agent version participation over time

Agent Version Distribution (24h)

Agent Version Distribution (24h)

Shows the versions of one agent type (client software) among the online peers. One value is the average number of online peers per crawl in the 24 hours up to the latest crawl. Select the agent type in the menu. The menu lists the six agent types with the most peers, and the first one is the default. The eight most common versions are shown separately. Other contains all other versions and the peers where we cannot read a version from the agent version. The percentages are shares of all peers of the selected agent type.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Snapshot of versions active per agent within the last 24hrs

Availability

We categorize IPFS Amino DHT Server nodes with regard to their availability as follows:

  • “Online”: the node has been found online for 80% of time or more.
  • “Mostly Online”: the node has been found online between 40%-80% of time.
  • “Mostly Offline”: the node has been found online between 10%-40% of time.
  • “Offline”: the node has been found online less than 10% of time.

Availability Over Time

Availability Over Time

Shows the number of DHT server peers per week in the last year, split by how long each peer was online in that week. One bar is one week, and a week starts on Monday at 00:00 UTC. Our Nebula crawler finds the peers in the DHT and tries to connect to them in every crawl. From these crawls we know how much of the week each peer was online:

  • Online: online for 80% of the week or more.
  • Mostly Online: online for 40% to 80% of the week.
  • Mostly Offline: online for 10% to 40% of the week.
  • Offline: online for less than 10% of the week, or never reachable.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

IPFS Amino DHT Server node availability over time

Availability Current Week

Availability Current Week

Shows the DHT server peers of the most recent complete week per continent, split by how long each peer was online in that week. A week starts on Monday at 00:00 UTC. Our Nebula crawler finds the peers in the DHT and tries to connect to them in every crawl. From these crawls we know how much of the week each peer was online:

  • Online: online for 80% of the week or more.
  • Mostly Online: online for 40% to 80% of the week.
  • Mostly Offline: online for 10% to 40% of the week.
  • Offline: online for less than 10% of the week, or never reachable.

The continent comes from the IP address mapping of ipregistry.co. Unknown contains the peers without a known continent.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

IPFS Amino DHT Server node availability - per region - over the current week

IPFS Amino DHT Server Peer Dynamics

IPFS Amino DHT Server Peer Dynamics

Shows how many DHT server peers join and leave the IPFS Amino DHT per day. Peers Joining counts the peer IDs that our Nebula crawler found online on a day but not on the day before. Peers Leaving counts the peer IDs that were online on the day before but not on that day. A peer is online on a day if it was online in at least one crawl of that day.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Aggregation of server peers entering and leaving the IPFS Amino DHT per day

Protocols & Infrastructure

Key metrics on supported protocols per agent version, as well as intrastructure providers.

Quic Support Over Time

Quic Support Over Time

Shows the number of online peers per crawl by QUIC support. The areas are stacked, so their top is the number of all online peers. A peer supports QUIC if at least one of its known addresses is a QUIC address:

  • QUIC/IPv4 + QUIC/IPv6: QUIC addresses over IPv4 and over IPv6.
  • QUIC/IPv4 only and QUIC/IPv6 only: QUIC addresses over one IP version only.
  • No QUIC support: no QUIC address.

Select an agent type in the menu to see only its peers.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Count of nodes that support QUIC per agent version

Cloud Provider Dependency Over Time

Cloud Provider Dependency Over Time

Shows the number of online peers per cloud provider in each crawl. The areas are stacked, so their top is the number of all online peers. The providers with the most peers in the selected time range have their own series. The other providers are in Other. Non-Cloud contains the peers that do not run in a cloud data center and the peers without IP address data. The cloud flag and the provider of an IP address come from ipregistry.co.

Select an agent type in the menu to see only its peers.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Count of nodes deployed per cloud provider and per agent

Supported Protocols Over Time

Supported Protocols Over Time

Shows the number of online peers per crawl that support a protocol, for the 40 protocols with the most peers in the selected time range. A peer counts once for each protocol that it supports, so the lines do not add up to the number of peers. The protocol list of a peer comes from the libp2p identify exchange. Peers without a known protocol list are not in this chart. Select an agent type in the menu to see only its peers.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Number of peers supporting the top 40 protocols in the network.

Supported Transports

Supported Transports

Shows the number of online peers per transport. One value is the average number of online peers per crawl in the 24 hours before the latest crawl. The transports come from the addresses of the peer, for example TCP, QUIC, WebSocket, WebTransport, or WebRTC. A peer counts once for each transport, so the bars can add up to more than the total. The dashed line shows the total number of online peers, and the percentages are shares of this total. Select an agent type in the menu to see only its peers.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Network snapshot in terms of supported transports

Supported Transports Over Time

Supported Transports Over Time

Shows the number of online peers per transport in each crawl. The transports come from the addresses of the peer, for example TCP, QUIC, WebSocket, WebTransport, or WebRTC. A peer counts once for each transport, so the lines can add up to more than the total. The dashed line shows the total number of online peers. Select an agent type in the menu to see only its peers.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Count of nodes support per transport protocol

Errors

Connection and crawl errors breakdown

Dial Error Causality Distribution

Dial Error Causality Distribution

Shows the share of the peers in each crawl by the result of the dial, the connection attempt of our Nebula crawler. One value is a percentage of all peers that the crawler tried to reach in that crawl. Each peer counts once. Peers that we count as online are in no_error. For the other peers, the chart shows the error that occurred most often on their addresses. no_public_address means that the peer had no public address to dial. The no_error series is hidden by default: click it in the legend to show it. Then all series add up to 100%.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Percentage split of dial errors per error cause

Crawl Errors Over Time

Crawl Errors Over Time

Shows the share of the peers in each crawl by the error that our Nebula crawler got when it asked the peer for its routing table. One value is a percentage of all peers that the crawler visited in that crawl. Only peers that accepted a connection can have a crawl error. All other peers are in no_error, also the peers that the crawler could not connect to. This series is hidden by default: click it in the legend to show it.

Data is provided as is, without warranty. Scraping and automated bulk extraction are not permitted. See our Terms of Use.

Percentage split of crawl errors per error cause

Related Work

Blog posts, papers and talks by the ProbeLab team about IPFS.

og-libp2p
Blog post

Can libp2p punch through NATs?

How efficient is libp2p’s DCUtR protocol in connecting actual peers directly?

IMC '26 Conference Paper
Paper

Large-Scale Measurement of NAT Traversal for the Decentralized Web: A Case Study of DCUtR in IPFS

The promise of decentralized peer-to-peer (P2P) systems is fundamentally gated by the challenge of Network Address Translation (NAT) traversal, with existing solutions often reintroducing the …

og-ipfs
Blog post

Optimistic Provide: How We Made IPFS Content Publishing 10x Faster

How a statistical approach cut IPFS content publishing times by over one order of magnitude.

IPFS Logo Blog Hero
Blog post

IPFS Gateway Performance Measurements

The ProbeLab team has been busy building new IPFS Gateway performance measurements

From rawpixel.com / Freepik
Blog post

ProbeLab's 2025 Review and 2026 Outlook

What we've achieved in 2025 and what we're targeting in 2026

University Other
Paper

Measuring and Optimizing Peer-to-Peer Networks: A Case Study of the InterPlanetary File System

The Internet's transformation from a decentralized network of networks into a landscape dominated by centralized platforms has created systemic risks including infrastructure fragility, censorship vulnerabilities, …