Bitcoin Forum
October 31, 2026, 07:05:12 PM *
News: Latest Bitcoin Core release: 32.1 [Torrent]
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: Bitcoin node stats from ProbeLab: IPv4, IPv6, Tor and I2P (bitnodes alternative)  (Read 15 times)
probelab (OP)
Newbie
*
Offline

Activity: 1
Merit: 0


View Profile
October 29, 2026, 08:16:15 PM
 #1

Hi all,

We're ProbeLab. We measure peer-to-peer networks like IPFS, Ethereum and Filecoin, and since a few weeks we are also crawling Bitcoin (and derivative networks like Litecoin, Dogecoin, Bitcoin Cash, and more). The results are public at probelab.io/bitcoin. We'd like feedback from people here, especially where something looks off.

Methodology
Our crawler starts from the DNS seeds, sends getaddr to every node it reaches, and dials every public address it gets back, over IPv4, IPv6, Tor and I2P. A node counts as online if it completes the version handshake during the crawl. Each address counts once, so a node that listens on IPv4 and Tor counts twice. A full crawl takes around 17 minutes. The latest one found about 29,200 online addresses.

If you're used to Bitnodes, expect different totals. Going by its public code, Bitnodes only dials addresses with a recent timestamp (the window moves between 8 hours and 5 days), and it doesn't dial I2P. We dial everything nodes hand out. No node gets skipped because its address looks old. Both approaches make sense. They just measure slightly different things.

Brief, incomplete list of things we show:
  • Sync state: each node's height compared to the majority tip. Nodes are in sync (within 12 blocks), behind, a week or more behind, in IBD, or ahead and probably on another chain. We spot IBD because Core sends the maximum fee filter while it syncs.
  • Minimum relay fee: how many nodes accept 0.1 sat/vB (Core 28.3, 29.1 and later) and how many still want 1 sat/vB, read from the feefilter message each node sends.
  • Suspected fleets: some nodes report height 0, never answer getaddr, and send no fee filter or broken messages. These are probably not real full nodes. We group them by fingerprint and by ASN. They stay in the totals, but you can see how many there are and where they're hosted.
  • Feature adoption: BIP324 v2 transport, compact block filters, Erlay, pruned and blocks-only nodes, over time and by client version.

There are also the usual client, version, country and ASN charts, including how few ASNs host half of the network. The crawler connects with the user agent /nebula:<version>/ and disconnects once it has the address reply. If you want your node left out, just filter for the above string. More information at probelab.io/abuse.

Questions about the method are welcome. If a chart doesn't match what you see on your own nodes, tell us and we're eager to learn. Also if you have an idea how to extend the service. We're also happy to hear suggestions!

Cheers
Dennis from ProbeLab
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!