In preparation for potential improvements to the Bitcoin reference client, citizens of Bitcoin's Most Serene Republic have begun establishing high availability Bitcoin nodes located in datacenters and hosted on dedicated hardware to serve as publicly advertised seed nodes from which Bitcoin users may sync their blockchains and more peers. The improvements to the foundation's reference implementation which spurred this move are the removal of all Domain Name System code from the client to allow an entirely statically built bitcoind as well as the removal of the legacy IRC initial peer discovery mechanism to reduce the attack surface bitcoind presents.
The experimental bitcoind recipe named 'stator'1 was submitted by Stanislav Datskovskiy to the BTC-dev mailing list on June 28th, 2015 and in the intervening period a number of the Republic's citizens have built the stator bitcoind on a variety of Linux systems ranging from Gentoo, through Red Hat, and all of the way to Mint and other Ubuntu like Linux distributions. There are also people successfully building stator on OpenBSD, and building this bitcoind should be possible on nearly all sane Unix systems.
The removal of DNS however means that before stator bitcoind can run for the first time it requires a bitcoin.conf file with at least one reachable node specified through the "addnode" or "connect" parameters, hence the Republic's move to begin publicly advertising the existence of highly available Bitcoin nodes capable of supporting many connections. Once bitcoind has been started at least once it will have collected further peers it can use to continue its commune with the Bitcoin network. The first two of The Republic's advertised seed nodes are:
- 188.68.240.167:8333 Supported by Mircea Popescu and advertising itself as a Bitcoin version 0.6.2 node
- 195.211.154.159:8333 Supported by No Such lAbs, who also offers the Phuctor RSA Super Collider service. This particular node advertises its protocol version as 99999 which makes it visible on Bitnodes.io2
As the stator improvements undergo further testing and approach approval for a Bitcoin Foundation release, the number of seed nodes provided by Republic Citizens is likely to grow.
A reference to the static part of a rotary system ↩
Bitnodes insists they only show nodes that are protocol version 70001 or newer, with the intention of hiding all pre-version 0.8.x nodes. Because the check is only of a self reported version number, it can be counted for as much as any other self reported thing. ↩
Despite forcefully connecting to both seed nodes, I get no other pre-0.8.x node (nor did I when I wasn't doing it, over hundreds of nodes).
Top versions are /Satoshi:0.10.2/, /Satoshi:0.9.2.1/, /Satoshi:0.10.1/
There is however a surprising amount of BitcoinJ.
Same story from other nodes: https://wallet.la/connectednodes http://87.104.113.2:8334/peers (found those by searching for "UMDCoinscope" as I wanted to find out what it was).
Don't expect to find any which explicitly advertise pre-0.8.x. Just about everyone did the same trick as pictured here – tweak the version string to avoid being easily filtered by Gavin and his many merry friends.
But there's no filter. In fact, I see random user agents. It's awfully convenient to claim people bother masquerading as new Bitcoin Core versions. More than that actually, that *everyone* running an older version is masquerading as a new version!
https://bitcoin.toshi.io/#block
Available peers
17,642
Number of peers discovered by toshi
It may be written in Ruby, but toshi finds a lot of peers. Wonder what versions they are.
That's a shame. I would have rather seen who actually uses it. Damn. Shoulda been pretty obvious if they start filtering – and you could use a distinct name rather than 0.5.x (i.e "most serene version 0.5.x")
> 188.68.240.167:8333 Supported by Mircea Popescu
bravo.
> and advertising itself as a Bitcoin version 6.2 node
I think you mean 0.6.2? If so, bravo again.
gbangalangadingdong
Thank You.
Are you Romanian, muh prince ?
Could you confirm the minimum relay fee for these nodes? I tried 0.0009 and it worked (with the S.NSA one). However 0.0001 was too little, apparently. Are they set to 0.0005 and can we expect it to stay like that for the immediate future?
It's interesting to know, so one can ask the sender to always set fee at least to that amount, if they want us to see their transactions immediately.