[ANN] BitEN - bitcoin erlang node

[r.willis]

Hello all!
For some time I was working on bitcoin daemon implementation in erlang.
Now it's in the early stage of development. I'm trying to implement protocol not looking into bitcoind source, but using specification from https://en.bitcoin.it/wiki (found one bug so far), original paper and looking at the protocol messages which I get from peers.

My goal is to make scalable bitcoin daemon, capable of serving many thousands of peers on relatively low-end servers.
At current state it's capable of connecting and parsing messages from 3000+ peers on 256M RAM VPS without swap. It does not do much besides protocol parsing and peer discovery now, but it's only several days old. Chances are that it connected to your machine - check for /ebtcd:0.0.1/ user-agent in logs.
Short term pans: accept incoming connections, announce own address via addr message.
Mid term plans: add validation and routing (test on testnet), release source under open license (AGPL sounds good).
Long term: full-featured daemon.

P.S. In case someone wants to encourage me with some BTC, I set up donation adress:
1GVmS56pvVL7YZA7YqMBXmaDedCoputKuJ

UPD:
git clone git://github.com/r-willis/biten.git

[1714079801]

1714079801

[1714079801]

1714079801

[1714079801]

1714079801

[gweedo]

Any chance you going to put it on gihub?

[kokjo]

please make the source public, og remove [ANN] from the thread subject... i was expecting to see awesome code, but was disappointed.

[r.willis]

please make the source public, og remove [ANN] from the thread subject... i was expecting to see awesome code, but was disappointed.

You are right, so [ANN] gone for now.
Code is in no way awesome now. It's a bunch of loosely connected modules and processes. Current code is proof-of-concept that you can make "very-well-connected" node on low-end virtual server. Right now I'm converting code to OTP (Erlang framework for robust, industrial-grade applications).
I'd expect code release in a week or two.
Why I (we) want node with many peers? Because it greatly reduces network "diameter", reduces propagation times, helping miners and makes many attacks not feasible.
But if it does not verify transactions/blocks it relays, I effectively becomes threat to network, as if there is many such nodes, spammers can use them to amplify their abilities to flood nodes with malformed/fraudulent transactions.  

[kokjo]

But if it does not verify transactions/blocks it relays, I effectively becomes threat to network, as if there is many such nodes, spammers can use them to amplify their abilities to flood nodes with malformed/fraudulent transactions.  

i don't think so... the satoshi client have some ban system, if your node began to send malformed txs they would just disconnect from your node.

[grau]

Now it's in the early stage of development. I'm trying to implement protocol not looking into bitcoind source, ...

From this I can tell that this must be really early stage. Keep up, but be prepared for a very long trip.
Unfortunately documentation available is very far from being sufficient input for an implementation, and in Bitcoin world there is no half-way compatibility.

[r.willis]

i don't think so... the satoshi client have some ban system, if your node began to send malformed txs they would just disconnect from your node.

Of course it has, but it but before it kicks in, you can spend several megs of some someone's bandwidth. Multiply by some thousands peers, and by number of servers (whatever it is). I just not very comfortable with this.

Keep up, but be prepared for a very long trip.

Yes, I read your thread about bitsofproof. I hope it will be interesting journey.

[kokjo]

i don't think so... the satoshi client have some ban system, if your node began to send malformed txs they would just disconnect from your node.

Of course it has, but it but before it kicks in, you can spend several megs of some someone's bandwidth. Multiply by some thousands peers, and by number of servers (whatever it is). I just not very comfortable with this.

and you should not be, but its still safe to experiment, the bitcoin network is robust(at least the satoshi client is...)

i am also devloping a bitcoin client, pycoin. although i have not done much lately on it.

[r.willis]

If you think so, I can implement routing and release code. Let's stress-test network.
I think on 8Gb (right now, without optimisations) machine daemon can connect to every connectable node in network (I estimate it's around 60k peers) + some incoming connections.

[jgarzik]

i am also devloping a bitcoin client, pycoin. although i have not done much lately on it.

If you like python, check out https://github.com/jgarzik/python-bitcoinlib/ and the bitcoin client that uses it, https://github.com/jgarzik/pynode/

I'm trying to encourage everyone to use python-bitcoinlib for their python projects.  It fully verifies the blockchain, and passes many reference client tests.

[meowmeowbrowncow]

i am also devloping a bitcoin client, pycoin. although i have not done much lately on it.

If you like python, check out https://github.com/jgarzik/python-bitcoinlib/ and the bitcoin client that uses it, https://github.com/jgarzik/pynode/

I'm trying to encourage everyone to use python-bitcoinlib for their python projects.  It fully verifies the blockchain, and passes many reference client tests.

And that Python, generally, is cool.

[aantonop]

Very interesting.

You should open source even if it is early stage, please! I'd love to

I've done some Erlang and OTP and I'm looking at core-daemon (full node) implementations.

Don't be shy, open it!

[kokjo]

i am also devloping a bitcoin client, pycoin. although i have not done much lately on it.

If you like python, check out https://github.com/jgarzik/python-bitcoinlib/ and the bitcoin client that uses it, https://github.com/jgarzik/pynode/

I'm trying to encourage everyone to use python-bitcoinlib for their python projects.  It fully verifies the blockchain, and passes many reference client tests.

i have been looking at thats python codes before. but it does not seem 'pythonic', but the script parts of the code i have found helpful.

[r.willis]

Conversion to OTP mostly done.
Plans for tomorrow: limit rate at which outgoing connections are created. After this, accept incoming connections and advertise own address via addr message.

[kokjo]

Conversion to OTP mostly done.
Plans for tomorrow: limit rate at which outgoing connections are created. After this, accept incoming connections and advertise own address via addr message.

and then: github

[r.willis]

Before github, I need cool name for project. Any ideas?

[kokjo]

Before github, I need cool name for project. Any ideas?

"The first and only true bitcoin erlang deamon" or TFAOTBED?

[meowmeowbrowncow]

Before github, I need cool name for project. Any ideas?

"The first and only true bitcoin erlang deamon" or TFAOTBED?

How about:

The Committee for the Liberation and Integration of Terrifying Organisms and their Rehabilitation Into Society




CLITORIS

(had to, sorry )

[r.willis]

It's hilarious... I'll consider it.

[r.willis]

System is live at 192.210.207.147:8333
Accepts connection and does handshake. Nothing fancy yet. Transaction routing and block download/serving are next.