The following charts on the Bitcoin Peer-to-Peer Network are based on measurements conducted for research purposes. Our monitor nodes are located in Germany, results may differ for other locations.

Connection Count

Change plot range: day week month year all

Description: We operate two monitor peers that try to establish connections to all reachable peers of the bitcoin P2P network. The number of connections includes IPv4 and IPv6 addresses. As peers may be dual stacked (i.e., reachable via IPv4 and IPv6), the total number of unique reachable peers is most likely lower than what the plot suggests. The total size of the bitcoin network, however, also includes non-reachable peers and is therefore much bigger.

Churn (per Hour)

Change plot range: day week month year all

Description: Here, churn means how many connections are established or closed per hour to each of our monitor nodes. We do not know how many peers actually join or leave the network. Peaks in this plot are most likely caused by an interruption of our internet connection.

Distribution of Connection Durations

Change plot range: day week month year all

Description: This plot shows the share of peers connected to our monitor nodes for at least a certain amount of time. E.g., More than 90% of peers are connected for at least one hour. Again, we do not know how long other peers remain in the network, only how long they are connected to our peers.

Transactions per Hour

Change plot range: day week month year all

Description: This plot shows the number transactions broadcasted through the bitcoin P2P network per hour. No blockchain information is used here (i.e., it does not matter whether a transaction is included in a block or not). Transactions that allow linking the transaction to an IP address that first broadcasted the transaction in the network are separately given. Please note that this linking does not imply that a transaction was also issued by the linked IP address.

Current Propagation Delay Distribution

Description: The elapsing time between the first reception of a transaction and the subsequent receptions from other peers is shown here. This plot does not show when network peers receive transactions, only when they send transactions to our monitor peers.

Transaction Propagation Delay History

Change plot range: day week month year all

Description: Based on above data we calculated the time it takes until 50% (90% resp.) of network peers sent a specific transaction. After the acceleration of information propagation in late 2015, propagation was slowed down again by the release of Bitcoin Core version 0.12 in February 2016. Gaps in the measurements were caused by data loss.

Block Propagation Delay History

Change plot range: day week month year all

Description: Based on above data we calculated the time it takes until 50% (90% resp.) of network peers sent a specific block.

Current Activity on the Bitcoin Network

Description: A considerable number of peers on the bitcoin P2P network do not relay transactions or blocks to our monitor peer. We calculate an activity score for each connected peer by dividing the number of INV messages received from the peer by the total number of unique INV messages observed. An activity score of 1 means that the peer announced all possible transactions and blocks to us, an activity score of 0 means that the peer did not announce any transaction or block.

Change plot range: week month year all

Description: Based on above data the number of peers with an activity score of less than 10% is shown, contrasted by the total number of peers.

Research

The results presented here were measured using the infrastructure originally set up for this work:
Timing Analysis for Inferring the Topology of the Bitcoin Peer-to-Peer Network (Presentation)
@inproceedings{neudecker-atc16,
  title = {Timing Analysis for Inferring the Topology of the Bitcoin Peer-to-Peer Network},
  author = {Neudecker, Till and Andelfinger, Philipp and Hartenstein, Hannes},
  booktitle = {Proceedings of the 13th IEEE International Conference on Advanced and Trusted Computing (ATC)},
  year = {2016},
  month = {July}
}
Furthermore, the gathered data was used for this work:
Could Network Information Facilitate Address Clustering in Bitcoin? (Presentation)
@inproceedings{neudecker-fc17,
  title={Could Network Information Facilitate Address Clustering in Bitcoin?},
  author={Neudecker, Till and Hartenstein, Hannes},
  year = {2017}
}
If you have any questions regarding measurement methodology, raw data, etc., feel free to contact us!

Other Bitcoin Monitoring Sites

P2P Network

Bitnodes
Bitcoin Stats

Other (Mining, etc.)

Network Hashing Rate
blockchain.info
and many, many more...

Misc

This work was supported by the German Federal Ministry of Education and Research (BMBF) within the project KASTEL IoE in the Competence Center for Applied Security Technology (KASTEL) and performed by the Decentralized Systems and Network Services Research Group at KIT.

Contact