Are there enough nodes online?

[slob]

I was looking at the list of nodes connected to blockchain.info

http://blockchain.info/connected-nodes

It says a total of 706 nodes are connected. Does that mean, right now, there are only 700+ copies of the blockchain online?

[1715338823]

1715338823

[Kris]

Connected to blockchain, most likely. However each of the nodes listed could be connected to 706 other nodes, and each of these 706 other nodes could be connected to 706 other nodes.

[TippingPoint]

And now...

Total Connected: 926

[eleuthria]

I was looking at the list of nodes connected to blockchain.info

http://blockchain.info/connected-nodes

It says a total of 706 nodes are connected. Does that mean, right now, there are only 700+ copies of the blockchain online?

It means there were only 700 nodes *connected to blockchain.info*.  There are significantly more nodes than that in total.

[franky1]

I was looking at the list of nodes connected to blockchain.info

http://blockchain.info/connected-nodes

It says a total of 706 nodes are connected. Does that mean, right now, there are only 700+ copies of the blockchain online?

no this means that there are 706 nodes talking directly to blockchain.info.

each of those nodes also talk to other nodes.

here is a better stats page for the entire network of nodes instead of just the ones directly connected to blockchain.info
http://getaddr.bitnodes.io/

[No_2]

Whilst looking at a payment address on blockchain.info on a colleague's computer I sent a payment to him from my mobile phone wallet.

The payment registered on his's screen as I was hitting the button on my phone, it must have been under 100 milliseconds. Very very fast propagation time.

One understands these things logically, but seeing it is very reassuring from an intuitive perspective.

[TippingPoint]

Whilst looking at a payment address on blockchain.info on a colleague's computer I sent a payment to him from my mobile phone wallet.

The payment registered on his's screen as I was hitting the button on my phone, it must have been under 100 milliseconds. Very very fast propagation time.

One understands these things logically, but seeing it is very reassuring from an intuitive perspective.

I was also surprised when I saw that happen the first time.
https://bitcointalk.org/index.php?topic=137272.msg2227831#msg2227831

[TippingPoint]

 The top 10 countries with their respective number of active nodes are as follow.
1   United States   31124
2   China   15658
3   Germany   10602
4   United Kingdom   7526
5   Russian Federation   7484
6   Canada   4890
7   Australia   3671
8   Poland   2557
9   Ukraine   2505
10   Netherlands   2488

[drawingthesun]

bitnode says less than 200,000 nodes are in the world. This is the most depressing thing I have heard in a long time.

I expected several million active nodes at least.

[jubalix]

bitnode says less than 200,000 nodes are in the world. This is the most depressing thing I have heard in a long time.

I expected several million active nodes at least.

200000 are quite a lot considering, I mean how many companies/govs/banks can afford 200K backups of the same data

 

[slob]

200,000 copies of the blockchain at 10gb a pop— that's like 2 petabytes!

[jgarzik]

Answer: no.

We can always use more full nodes (bitcoind or Bitcoin-Qt) that accept incoming connections from the global Internet.

If you are behind a firewall (you should be!), drill a hole for port 8333, and verify that it works from somewhere outside your network.

[gmaxwell]

bitnode says less than 200,000 nodes are in the world. This is the most depressing thing I have heard in a long time.
I expected several million active nodes at least.

that number is pure fantasy. It's culled from addr broadcasts, but there are crap nodes that just send random addresses to anyone who connected to them, greatly inflating estimates based on watching addr broadcasts. (and at the same time, there are a great many nodes which never addr broadcast)

There are on the order of 4000 to 5000ish listening nodes. Average connection counts on listening nodes suggest that there are something on the order of 50,000 total bitcoin nodes.  It's likely that we're soon going to be going through cycles of running out of listening sockets again, this happened back in early 2011 due to rapid growth and now due to greater numbers of people running bitcoinj (android wallet, multibit, etc.) which consumes connections but doesn't contribute back to the network.

A lot of people use Bitcoin only via websites, what this says for the future security of bitcoin is hard to say.

[drawingthesun]

There are on the order of 4000 to 5000ish listening nodes. Average connection counts on listening nodes suggest that there are something on the order of 50,000 total bitcoin nodes.  It's likely that we're soon going to be going through cycles of running out of listening sockets again, this happened back in early 2011 due to rapid growth and now due to greater numbers of people running bitcoinj (android wallet, multibit, etc.) which consumes connections but doesn't contribute back to the network.

A lot of people use Bitcoin only via websites, what this says for the future security of bitcoin is hard to say.

Would a small monetary incentive towards being a full node help matters much? It looks like running a node actually offers no advantage and plenty of disadvantages to online and lite nodes.

[OnkelPaul]

Would a small monetary incentive towards being a full node help matters much? It looks like running a node actually offers no advantage and plenty of disadvantages to online and lite nodes.

I used to start up my bitcoin node (incoming connections enabled) with my computer at home. Stopped doing that when I noticed that my complete outgoing network bandwidth was consumed by bitcoin traffic, noticeably affecting other network uses.
Since I don't trigger many transactions, I now only run the client once or twice a week to catch up with the blockchain.

Onkel Paul

[Akka]

There are on the order of 4000 to 5000ish listening nodes. Average connection counts on listening nodes suggest that there are something on the order of 50,000 total bitcoin nodes.  It's likely that we're soon going to be going through cycles of running out of listening sockets again, this happened back in early 2011 due to rapid growth and now due to greater numbers of people running bitcoinj (android wallet, multibit, etc.) which consumes connections but doesn't contribute back to the network.

A lot of people use Bitcoin only via websites, what this says for the future security of bitcoin is hard to say.

Would a small monetary incentive towards being a full node help matters much? It looks like running a node actually offers no advantage and plenty of disadvantages to online and lite nodes.

That's one of the disadvantages of Bitcoin and what I think it's one of its most serious flaws. Bitcoin pays you for producing hot air, but there is no financial inactive in providing the resources it actually needs. Bandwidth and Space. In the original design this was tied together as in order to mine you also needed to run a full node. In this design every miner would, besides buying high priced electric heaters, need to invest money into having a well connected node and to store the Blockchain.

 The invention of pooled mining changed this. Now Mining Pools are the only ones that rally need to run full nodes. Some Business maybe will also invest in nodes, but it isn't absolutely necessary for them.

[gmaxwell]

Would a small monetary incentive towards being a full node help matters much? It looks like running a node actually offers no advantage and plenty of disadvantages to online and lite nodes.

Crazy. It offers the advantage of having security that is independent of other people. Even though your 'online' or lite node may keep its keys privately, this doesn't prevent parties you're connecting to from doing thing like lying to you about payments you've received (at some costs for true SPV nodes).  Even independent of the personal security benefits, it contributes to the health of the network backing these coins that you own.

That's one of the disadvantages of Bitcoin and what I think it's one of its most serious flaws. Bitcoin pays you for producing hot air, but there is no financial inactive in providing the resources it actually needs. Bandwidth and Space. In the original design this was tied together as in order to mine you also needed to run a full node. In this design every miner would, besides buying high priced electric heaters, need to invest money into having a well connected node and to store the Blockchain. The invention of pooled mining changed this. Now Mining Pools are the only ones that rally need to run full nodes. Some Business maybe will also invest in nodes, but it isn't absolutely necessary for them.

Mining is something the network needs, don't understate that— but indeed,  we have frighteningly few actual miners in bitcoin a couple dozen pools plus p2pool users, pretty much.  Everyone is is just selling computing power to one of those parties. It's fragile and frightening.

It's possible to have a POW scheme which prevents this, but the tradeoffs aren't obvious, alas.

[drawingthesun]

Mining is something the network needs, don't understate that— but indeed,  we have frighteningly few actual miners in bitcoin a couple dozen pools plus p2pool users, pretty much.  Everyone is is just selling computing power to one of those parties. It's fragile and frightening.

Is it true that Satoshi never foresaw pooled mining? Did he ever comment on the fact that there will eventually only be a handful of miners in the world?

It's possible to have a POW scheme which prevents this, but the tradeoffs aren't obvious, alas.

Could you give me a idea what would work?

[gmaxwell]

Is it true that Satoshi never foresaw pooled mining? Did he ever comment on the fact that there will eventually only be a handful of miners in the world?

I don't believe that he did, it's kind of crazy when you think about it. It undermines the purpose and motivations of creating Bitcoin. If you don't actually care about decentralized cryptocurrency, why are you doing anything with Bitcoin in the first place? ... Because it's profitable, but that were so obvious it would have been created long before Satoshi did.

With a little more foresight, Satoshi could have both foreseen pooled mining and possibility of using the same consensus algorithm to decenteralize pooled mining (e.g. P2Pool) and perhaps the symmetry would have been broken differently, large centralized pools might not have arisen, and the world might be a very different place now. Alas.

Could you give me a idea what would work?

See: "POW which involves queries against the UTXO set"

A whole bunch of variations are possible, but the general idea is to use the state you're mining against to make a deterministic random selection in a normative UTXO set, and then hash the result and thats your POW.  So your mining hardware would be boards of ultra fast flash memory coupled with sha256 fpgas or something instead of SHA256 chips.  It's a kind of memory hard POW.

[jubalix]

Is it true that Satoshi never foresaw pooled mining? Did he ever comment on the fact that there will eventually only be a handful of miners in the world?

I don't believe that he did, it's kind of crazy when you think about it. It undermines the purpose and motivations of creating Bitcoin. If you don't actually care about decentralized cryptocurrency, why are you doing anything with Bitcoin in the first place? ... Because it's profitable, but that were so obvious it would have been created long before Satoshi did.

With a little more foresight, Satoshi could have both foreseen pooled mining and possibility of using the same consensus algorithm to decenteralize pooled mining (e.g. P2Pool) and perhaps the symmetry would have been broken differently, large centralized pools might not have arisen, and the world might be a very different place now. Alas.

Could you give me a idea what would work?

See: "POW which involves queries against the UTXO set"

A whole bunch of variations are possible, but the general idea is to use the state you're mining against to make a deterministic random selection in a normative UTXO set, and then hash the result and thats your POW.  So your mining hardware would be boards of ultra fast flash memory coupled with sha256 fpgas or something instead of SHA256 chips.  It's a kind of memory hard POW.

do you think this could be an innovation in CC's, forced decentralized mining, or does sunnykings proof of stake give another out?