Bitcoin Forum

In these times when nearly all indicators are pointing towards increasing adoption and growing economy of Bitcoin there is one important (??) indicator pointing down: The number of connected hosts.

According to http://bitcoinstatus.rowit.co.uk/ there are now less than 3000 listening hosts. I guess these are mostly people running Bitcoin-qt and with port 8333 open - mainly miners.

I do not think that this development is surprising: Most regular Bitcoin users are migrating to lightweight clients and mobile clients and only miners really need to remain well connected to the backbone p2p network.

In the beginning of Bitcoin -
anyone with a CPU could mine (one CPU - one vote) -
soon, only people with a GPU could mine (ruling out laptops) -
then, only people with certain ATI GPUs could mine (ruling out Nvidia and on-board graphics) -
today, only people with ATI GPUs AND cheap electricity or FPGAs could mine (ruling out many people who own even good ATI GPUs)
in the near future, only people with an expensive 28 nm FPGA can mine (even fewer people) -
within a year or so, only people with a dedicated ASIC (costing $1000+) can mine (even fewer people).

The thing is that while that the raw hashing power needed to attack the network is growing, the number of miners you need to target by DDOS or otherwise (if you are a malign government wanting to take down Bitcoin) is getting lower by the day!
IMO the 3000 nodes that we currently have seems like an easier target for a well funded attacker than than the prospect of deploying 55,000+ spartan 6 chips to brute-force a 51% attack.

What do you think? Is the chronic decline in hosts a genuine problem for the backbone p2p network?

It is still 1 'vote' per hash. I guess you are annoyed because so many people are so interested in bitcoin?

If you want a bigger say you could help the network more. If you don't know how you could hire someone. Lots of people selling hashing power lately.

Just looked at the graphs, more concerned about how many version 0.3 clients are still being used.

Most miners use pools.  I doubt there are 3000+ solo miners to begin with.

Looking at listening hosts is kinda silly.  Many bitcoind are configured to ignore incoming connections (most major pools are) so that stat really isn't meaningful.

Attacking the network by DDOS is one thing.  Trying to use it for double spend is another.  It is pretty obvious and users can simply stop moving funds to be "immune" to any effects.  If it becomes a problem I would expect all clients to include some "network health" metric and warn users to not spend/send coins when network is being disrupted.

I guess many people are running the bitcoin client through Tor, those don't show up.

OK, two questions:
1) Why is it not a problem that there are so few listening hosts? Would the network still function if there were zero?
2) What about the decline in total hosts? Is that not a problem for the network's ability to resist attack?

If listening hosts fell to 0 then it would not be possible for new nodes to boostrap.  That would be very bad.  The site you linked has a stat to indicate that danger.  I am not saying it is "good" just pointing out your cause and effect aren't well supported.

Total hosts declining does make the network more vulnerable to DDOS attack however that is academic at best.  The pools have long since represented 70%+ of hashing power.  DDOS the top 10 pools and you cut the effective hashrate significantly.  Still long term I don't think that will be viable.  If the top 10 pools went down I would expect most exchanges and major business to halt deposits/sales giving you no target for your 51%.

Keep in mind that 1 host is Deep Bit which is like 20-30% of the entire bitcoin network hash rate. Nuff said...

If the major pools were DDOSed, most miners would automatically switch over to backup pools, p2pool, or solo mining. I'd be more worried about the major pools getting hacked but even that wouldn't really hurt much of anything for very long.

I'm a miner but I only open the Bitcoin-QT once a day to download blocks and then I immediately close it.  It's open for approximately 90 seconds of any given day.

The number of nodes is not a accurate representation of the popularity or growth.
similar deal here. I only open it once a week to update it or to send transactions.

It looks like your upset because you cant mine and get what people think is free money?

I leave it always open and listening. I don't see a reason not to, having a plenty powerful pc and connection.

All 4 of my rigs are updated and always on. Both my laptops run a node each, and my younger brother and parents across town have a node on each of their computers. all but my primary have a 0 Balance, but they are there doing their jobs.

More and more personal nodes also are "quiet" nodes as well, whether or not they are listening hosts.  Basicly this means that they don't announce their IP address in any fashion, and may also primarily use a VPN, Tor or I2P.  Which means that they might accept incoming connections, but the incoming node must first be aware of the IP address of the quiet node.  This could be because those two nodes have connected in the past, or the IP address of the quiet node was revealled to the other node over the bitcon network itself.  Certainly not all nodes can be quiet, but any node can switch from quiet to normal to offline and back whenever the user wishes.  My own node is fairly quiet when I have it on, and does accept incoming connections from nodes that it recognizes, but isn't online at all most of the time.

How many connections do these nodes have?
When I just run bitcoin-qt I only have the 8 (outgoing) connections because the network (not under my control) does not allow free traffic on port 8333?

I suspect that at the current rate of adoption, it will become increasingly expensive to run a full client and there will be continued drop-out.

If, for some reason, Bitcoin became suddenly very popular, I suspect that the rate of drop-out will become very sharp.  Only the people with the ability to construct and operate 'super-nodes' will be able to do so.  At that point, a relatively small amount of arm-twisting at the corp/gov level and DOS will spell the end of Bitcoin-I as a reliable solution.

Of course it will be re-born but it will be much more difficult to convince people that p2p crypto-currencies are a solid solution.

If you truly feel that way, I'll be the adult an offer to relieve you of your entire burden of unspent bitcoins for $2 apeice.  Which, according to your theory, is $1.99 more than their true worth.

In the meantime, the rest of us are going to keep doing what were doing, okay?

This, +1

Not hardly.  I expect that prior to a collapse such as I've describe (as a possible scenario) BTC will shoot up in value.  I'll either capitalize then, or take a bet that controlling secret keys which are represented in what I call the 'Satoshi blockchain' will be even more valuable.  Probably I'll hedge my bets and do some of both.

I need BTC at around $100 per before I'm ready to sell my stash down to a nice magic number.  Ping me then if you are still interested in being a buyer.


Oh yes.  Absolutely!  Carry on as you will please!

MasterTroll™

I'm not trying to win any popularity contests here.  Obviously.

Trolling in my posts on this thread is actually fairly incidental.  I expect at most a handful of people to contemplate my conjectures.  Mostly I expect pom-pom wavers to do exactly as you'v done.

---

Rah! Rah! Rah!  Bitcoin is awesome and has no possibility of any defects whatsoever.  Keep stacking!

There.  Do you like me better now?

Me too. Plus, I also do this...


... since I found out some people are trying to link your IP to all your transactions.

Actually, I really think that putting bitcoin behind Tor should be a "recommended practice" as using different addresses for each transaction, and that allowing the bitcoin protocol to exchange, recognize and publish hidden services should be a development of high priority.
People should not be exposing all their transaction history so easily...

It is probably not prudent to expose port 8333 on an IP address behind which there are actual coins. If anything, people will start attacking that IP address in an attempt to get to your wallet. They may not get in but they will probably slow down your firewall/router especially if it is just consumer-grade.

I lease a couple of cheap $30/month VPS's to exclusively run bitcoind with maxconnections=200 and empty wallets. That way, I always have access to a couple of well-connected that I control and trust. My quiet clients (from where I do my real transactions, p2pool, etc) connect directly to these. Since other people also relay transactions through my public nodes, I also buy myself some anonymity with this setup. In fact, most of the transactions originating from my nodes are other people's.

There was the idea to also pay nodes with bitcoins.
I don't know if it was completely declined or just postponed

Just relaying through your public node is not the same as "originating" in it.
A hypothetical attacker connected to every, or nearly every bitcoin node would still see where the transactions actually originate.
Unless the "other people" you mention also do as you do, I mean, connect exclusively to your public node. But then they would need to trust you not to store their data.

I'm sorry, but how do you run a BTC client behind Tor? My Tor installation comes with a bundled browser, and from what I gather the only communications being scrambled are the ones the Tor browser makes... Am I wrong?

How would one (easily  :P) make MultiBit "speak" through Tor?

You aren't wrong, but you can route traffic through tor using proxy settings as seen here. It is easy.
https://en.bitcoin.it/wiki/Tor

On ubuntu listen it is as simple as installing tor, making sure it is started, and the adding proxy=127.0.0.1:9050 to your bitcoin.conf
 OR starting bitcoind with flag -proxy=127.0.0.1:9050

Don't know about windows/Mac, but it shouldn't be too hard... Tor is for encrypting/anonymizin all internet traffic, not just for web sites.

Ultra-cool-tip of the day: you can create a server, which is _only_ set up as a tor hidden service. This means that you can only ssh to it through tor, etc etc.

If your installation is working, you probably have a Tor-proxy running as a service in the background. By default, Tor SOCKS port is 9050, so normally all you have to do is configure your client to use as SOCKS proxy "localhost:9050".
Attention though: it's important that your client doesn't leak your identity through the communication, or by opening a listening port. Satoshi's client is designed to support Tor, so the devs take the necessary care for it not to happen. I suppose MultiBit's developers do the same, but you'd better ask them to be sure. (btw, that's why the tor people release a browser bundle, to make sure the browser does not leak your identity)

Actually no.  With fewer and fewer nodes using IRC for peer discovery, it becomes more and more difficult to track running nodes unless you're crawling them.  So basically, a bunch of improperly placed doom and gloom.

I'm surprised by the number of people saying, even though they maintain the block chain they prefer not to run their nodes continuously. Why?

(I'm not asking specifically about listening nodes though, I understand there can be concerns about announcing your IP.)


This doesn't sound right to me. How do you presume the attacks would take place? If they aren't trying to deny your connection, I'm not sure how it would have any effect. Is there a conventional attack that requires considerable amounts of bandwidth? How many different parties do we expect our listening node would be attacked by simultaneously?

Also, I wonder about the success rate of this kind of attacks in general. I run a listening node 24/7, and plan to continue doing so as long as the blockchain fits on a consumer-grade hard drive. I really don't expect attacks becoming a problem.


Good idea, but I hardly think such a setup can be deemed necessary.

It makes me sad that this this thread has become so long without anyone pointing our two important points.

(1) bitcoinstatus' results are broken, and have been broken for a long time— if not always

The code that tracks listening hosts for Pieter's dns seed is currently tracking 22,624 IPs, and 22,032 of them with uptime in the last 24 hours.

Has there been a decline in listening nodes?  Maybe. I certainly believe there has been a decline in total bitcoin nodes from the time when we had the wild popularity surge a year ago, but since then we've managed to get UPNP working correctly and enabled by default so I wouldn't be surprised to find out if we really had more listening nodes now than we ever had.

(2) there is no reason to assume those listeners are miners. There is no requirement to listen to mine, and in fact the higher relaying load and exposure to dos attackers can adversely impact mining— prudent miners separate those functions.

@ EhVedadoOAnonimato, BadBear, kangasbros: Thanks guys, I'll examine all the info you've given me ASAP!  8)

do they have ports open to the outside through your router?

Hey Guys... So I'm looking into all this right now, and I have a few more questions... I think I'll start a dedicated thead...

I know one attack involves trying to brute force the JSON password for clients (intentionally or accidentally) configured in server mode.  I can't recall if any countermeasures were added but if there aren't an attacker could generate a significant amount of bandwidth trying huge numbers of passwords.


I do to, although that is just to support the network my wallet is on a outgoing only node which connects to trusted peers.

Still, bruteforce would only apply to hosts that directly exposed 8332 not 8333. But having a hidden node talk to a trusted node is a good idea anyways.

Each is assigned to an external IP, if that is what you mean.

3 of my 4 miners are at various locations I work and play at. I pay any electric metered to them. my other rig, with my 5840 runs in my shed, with solar power and batteries in an inverter, its net connection is a tethered prepaid cell. My primary laptop uses a tethered cell for its internet and outgoing traffic, and my wifes laptop is on the home network, So thay all have unique External IP addresses.

I run a node when my pc is on, it's easy and fast, dunno why more ppl don't do that.

As I suspected, you have no earthly clue about the dangers that I was alluding to.  Oh well.

And in two years maybe anybody who can afford to buy an asic-mining-space-heater for their house can mine.  And in ten years maybe new houses will come pre-equipped with asic-mining-baseboard-electric-heaters.

Or maybe not.  Nobody knows what is going to happen, which is why I keep repeating BITCOIN IS AN EXPERIMENT IN PROGRESS.  If you can't handle uncertainty and the very real possibility that the experiment will fail, then don't get involved (but don't come back whining that you missed out if the experiment turns out to be a huge success).

No, I don't think it is a problem, any more than the relatively small number of "backbone" routers are a problem for the Internet. The bitcoin network is evolving in roughly the direction I expected it would (lots of lightweight clients connecting to a smaller number of heavyweight "backbone" nodes).

All of the incentives are for merchants, exchanges, and big mining pools to accept lots of connections, relay only valid transactions and blocks, and be as DoS-resistant as they can afford. Merchants and exchanges want transactions to be validated quickly, the big mining pools want their payouts to be processed quickly and want their blocks propagated quickly, and they all want to be reliable.  The big exchanges and mining pools have ALREADY implemented DDoS countermeasures because they have ALREADY been attacked. And the network keeps chugging away, processing transactions....

Thanks Gavin. Nice to know you also think that relatively few ddos-protected backbone nodes can provide a high degree of resistance to attack.
Sure, Bitcoin is an experiment, but some of us really want it to succeed since it has potential to cchange the world in a good way.

you trolling over here too, tvbcof?  ;D

I like to make friends wherever I go :)

Thanks for the explanation Gavin. As always, very concise.

I want Bitcoin to change the world too but I don't really understand how insulting the lead developer by implying he doesn't care if the project succeeds or not gets us closer to that goal.

Maybe that is the right approach to take and I'm just not seeing how it's beneficial to the project as a whole so can you explain it to me? I'll be more than happy join in and dump my emotional baggage all over the dev team if you can show me how that will make the project more successful.

I have a pretty significant difference of opinion here.  I'm sure it sounds paranoid, but I think that there is a more than slight chance that Bitcoin could be attacked by the infrastructure providers.

Before writing this possibility off completely paranoid ravings, remember back to the time when AT&T had no compunction about installing Narus's in spite of it being about as diametrically opposed to our 'principles' as one could get.  Yes, it seems like no big deal now, but that only shows how quickly things drift.  IIRC, the only telecom who resisted was Verizon and it's CEO became the target of interest to the US justice department (making it one of the very few cases that our justice department has taken an interest in any white-collar individual for about 15 years now.)  And, of course, all of the telecoms got blanket immunity for any past wrongdoing, and probably future wrongdoing as well.

Philosophically, I contend that Bitcoin is not much more than an interesting toy if it is not needed to replace other failing monetary solutions.  And if it is, it will be under significant attack and will be again not very useful unless it can withstand the attack.  I will bet money (by dumping BTC) that a 'a smaller number of heavyweight "backbone" nodes' will be much more susceptible to the kinds of attack which will pose a genuine threat.

So, what to do?  Honestly I have concluded that the answer is 'nothing much different'.

Taking some lessons from biology, one can try to kill a starfish by chopping it into 1000 pieces.  The net effect is you end up with 1000 starfish.  Even more impressive to me, one can smash a sponge through a strainer in an attempt to destroy it but it re-forms itself into a similar sponge on the other side.

While highly non-specific, I believe that things like the above mentioned 'outside the box' concepts and constructions are the kinds of models to look at in designing a robust and durable p2p crypto-currency solution.  Certainly Bitcoin has some novel, clever, and powerful concepts to protect it against attack, but it also has some 'tried and true' architectures who's weaknesses I stress about since, as they say, 'a chain is only as strong as it's weakest link.'  I think that it is not in the cards to rectify certain of these as a natural evolution of Bitcoin.

I suggest that the prime movers behind Bitcoin development, who I believe share the vision that crypto-currencies are a an empowering and positive thing for humanity, continue on the present course of making the solution as good as they can make it.  If/when a similar solution which has as a goal to address some of of the things which I've mentioned as concerns pops up, I do hope that there will be constructive symbiosis that will serve to strengthen everyone's hand.

You are falsely presuming trying to succeed == guaranteed success, it doesn't. There are no guarantees in life, Bitcoin is an experiment, thinking about it any other way will get you burned and all that Gavin is trying to do is warn you to stop it and take the appropriate precautions when using an experimental tool so you don't get burned unnecessarily.

I myself think if the total hosts are indeed declining, which I believe they are, this greatly diminishes Bitcoins resilience to central control and to being attacked by the state. The once touted peer to peer resilience similar to that of bittorrent is clearly slowly evaporating as time goes on and it would seem it's only a matter of time when we will have a few super nodes as the sole backbone and an easy target for either control or destruction.

Of course being an easy target doesn't mean it's worth to attack you. Bitcoin is still open source and still completely transparent and so an attack would with a high probability not go unnoticed meaning an attacker would have very little to gain facing the repercussions of their attack if their goal was control, and as long as a single copy of the blockchain is preserved and made freely available somewhere more people can start up a client if the need for defense should arise.

It's all very dynamic and fluid, it's praxiology of the human nature vs the integrity of an algorithm and technological advencment, how will it play out? No one knows.

Nice argument. So how else do we generate incentives for new nodes? I like the baseboard heaters idea  ;D

I think we would have a lot more nodes if we had a client that didn't need the whole blockchain and wasn't connecting to a server to get it. Personally I'm running the Satoshi client right now for the sole purpose of adding a node and I will continue to do so until I can afford the space.

Sipa has been busy implementing IPv6 support and much better support for running bitcoin as a Tor hidden service. Both of those should help make the network more robust and resistant to attack. If a lot of people decide to run only inside Tor then it will look like even fewer nodes are listening, though.

As a related side note...

At some point it will really not be very truthful to refer to Bitcoin as a 'p2p' solution.  It is on a trajectory which makes Bitcoin not much more 'peer to peer' than Visa, PayPal, etc (assuming they have some amount of clustering and redundancy.)

Of course even with the realization of Bitcoin as a 'necessary infrastructure backbone' architecture there will still be (hopefully) a lack of centralized control that we see in other solutions and that is a very good thing.  But it will be decidedly non-P2P so it might be worth thinking about how to shift various text and 'sales pitches' as the time approaches.  Or get beat up about false advertising, and legitimately so.

There's a distinction.  Anyone can run a bitcoin node of they want to, some are just opting to go the lightweight route.

I'm begining to suspect that you do not understand how bitcoin actually works, what decentralization means in the context of economic science, nor what it means to be a 'peer'.

When I log on to my Wells-Fargo account to transfer funds, is my computer a 'peer' to Wells's machines?

Until recently I did all my work on a machine which contained an up-to-date block chain and which I owned and controlled fully.  At that time I considered my presence in relationship to other participants in the economy as one of a 'peer', and I interacted with other participants directly (the '2' part) though I did rely on global network infrastructure which I pay a third party to provide.

_This_ is the archietecture which was described in most of the documentation related to Bitcoin, at least at the time I studied it significant.  As importantly, _this_ is one the aspect of Bitcoin which appealed to me the most about the solution.

and that's how it continues to be.  The people running lightweight clients CHOSE to run them.  Nobody forced them to run the lightweight clients.

no, he doesn't understand it.

To paraphrase our dear leader (in US-land and by extension in a lot of the rest of the world...):

  "I'm not interested in playing the game better; I'm interested in changing the rules of the game."

A big difference is that when Obama said it he was utterly and completely full of shit as history has shown.  But anyway, it looks to me like Satoshi also had similar instincts, and also that they were just a little bit to alien conceptually for a lot of folk.  Oh well.

Good one...

As long as anyone can join in, it's p2p. Doesn't matter who joins in. Of course you can't know for sure that anyone can join in until there is a critical number of peers present and there is a convincing level of political freedom (apparent conflicts of interest provides this). I don't think it's reasonable to expect it to naturally evolve to a state where it wouldn't be considered p2p.

No and if you have to ask something that dumb you likely shouldn't be making proclomaitons.

The very words "log on to my Wells-Fargo account" implies a master-slave relationship.

Your computer isn't joining the banking network.  Your computer is a dumb terminal.  Wells Fargo computer is part of the banking network.  A network you have no hope of EVER becoming a peer.  As such you are utterly beholden to the owners of that network to graciously give you an account, and allow you to make REQUESTS.   That is right, when you "transfer money" you aren't actually doing anything.  You can't do anything because you are an outside to the network.  You (via your dumb terminal browser) is formally asking Wells Fargo for permission, and if Wells Fargo feels like it they will grant your request.  If they don't (because they don't allow that, or you broke a rule, or the govt told them to stop) well then it becomes painfully obvious you aren't a peer and never ever ever will be.

If you can't see the difference with Bitcoin well ... never mind, outside of trolling it is impossible to not see the difference.



There is nothing preventing you from continuing that method of direct connection (unlike in your flawed banking analogy).  Now will the costs to run a node rise?  Maybe.  It is entirely possible that resource demands will grow but slower than Moore's law and as such the cost to you will remain the same (or fall).  Then again maybe you will incur a cost to have direct access to the network.

SO WHAT?

You have the Freedom.  Freedom is the choice.  It doesn't mean you will be given access free of charge.  Many people will choose the ease, convenience, (and potentially lower cost) of having indirect access.  As long as the number of peers remains robust and geo-diverse it isn't necessary for everyone to have direct access. 

Imagine a hypothetical situation where there are 20 million users but only 20,000 full nodes.  Who might make up full nodes?  Merchants, banks, exchanges, service providers, bitcoin enthusiasts, paranoid users, developers, and non-profits.  The fact that less than 0.1% of user's run a full node doesn't diminish the network.  At any point any one of them COULD run a full node.  That is the whole point.

Having full nodes easy to setup and run is a good idea, because it would means more people running them

I thought so too...which is why it is surprising that anyone would fail the recognize the rhetorical nature of the question.  Unless they had not really followed the thread and comprehended some of the conceptual and time-frame features of it.   The nature of your other comments indicate that this is what we have here.


Unfortunatly that, and vastly more, is quite easy to do.  And is the basis for a lot of my concerns...


And my point is that I am not very convinced that that will always be the case.  It could be, but I would consider Bitcoin a failure if so.  On it's present trajectory at least.

A better question is this...

Can you, without asking for special permission nor paying anyone for the privilege, set up a 'peer' to Wells' machines?  It's the ability to peer, not the reality, that makes bitcoin different from Wells Fargo.

Just getting up to speed on this topic, interesting issue.

Like so many here, don't keep my client-qt running all the time anymore, did first month or so, then stopped running it unless needed, try to update everyday but not always...  Summer time brings electrical storms and outages so prefer to keep the system off allot more of the time.

The problem is there is little incentive to do so.  If there was a way to gain some value, like free transactions for those clients left running or some type of payout dividend.  Not as much as mining, but something to support the contributors in maintaining a broad base of node participation.

Places like Australia, where upload bandwidth is charged every month on your bill, are even more discouraged to support the peer-to-peer paradigm.  Agree this trend down in number of available nodes is disturbing, and will lead to making the bootstrap of newcomers even more troublesome.   To bad those via Tor can't somehow be counted, be nice to have a more reviling graph of actual numbers...

There is no incentive for the end user to keep their bitcoind running 24/7 because there is no gain for the network, either.  A non-mining node that is just on the network listening and relaying adds nothing, so long as there exists somewhere one running with the entire blockchain.  It only takes one, and the mining pools' clients count.  Bitcoin does gain a little in the 'many-copies-keeps-data-safe' area, but that's just as true regarding a client that doesn't normally listen or one that just connects every few days to update it's blockchain.  This amounts to the greatest of non-issues for bitcoin.


Again, not relevent.  The very fact that the numbers quoted here can't be trusted should be proof enough that this issue isn't and issue.

To get back on topic; its unfortunate that the approximate number of bitcoin "users"  (full or lightweight hosts) in the last 24h or so can no longer be easily determined because at 3k and declining that stat can be used to predict the failure of bitcoin by its detractors.  Like any network technology, a lack of users discourages new users.  So if it is inaccurate, perhaps it should be removed (or relabeled) so as to not spread misinformation.  

And perhaps we should look for some other metric (or even proxy metric, such as # of IP address that read these forums) that would be more accurate so that if it is increasing we can use it to encourage new users.

This is a good point, and I respect that you have made it.  It is actually related to a part of my comment that you snipped out where I called attention to the very different nature of Bitcoin vs. others in terms of 'centralized control'.

Anyone who cares to do something other than wave the pom-poms may notice that I have a great deal of respect for and hopes for Bitcoin even in the 'worst' of circumstances and am certainly not immune to recognizing the strong points which it has.

We don't control that website, and thus don't control that metric.  Fudders are gonna fud.  It's what they do.

And the membership of this forum crossed 10K months ago, but there is no chance that the ownership of this forum is going to start counting IP addresses.  That wouldn't be any more relevant a metric anyway.  Not only do not all members run their own client, much less one full time; there are many more people who use bitcoin who don't have memberships.  Again, the growth or decline of the number of listening-but-not-mining clients is irrelevent to the function or resilence of the bitcoin network whether they are on the open Internet or some PVN.  Beyond some minimum number required to support the bandwidth of the network as a whole, that is.  The very fact that we can't know how many (or where are) all the network's nodes happen to exist is, itself, a contribution to it's resiliance.  An attacker can DOS the pools or exchanges, because he can find out it's IP address and a government agent can steal a server because he can find the farm that holds it; but these things cannot stop the bitcoin network for no other reason that you cannot kill what you cannot catch.  At worst, these kinds of events simply disrupt the network temporaroly and force more users towards Tor and I2P.

Fair enough.  Let's assume for a moment that your not trolling, and that your concerns are valid.

Do you have any suggestions for improving the protocol?

Hmm... that website doesn't really seem anti-bitcoin.  And if you're gonna write all that code you must be at least a little interested.  I was simply thinking someone here might know the owner and ask nicely. :-)


of course, but one can guess-estimate the number of non-forum members; just like radio stations estimate listeners from request phone calls... and is really the derivative that matters not the value anyway.


I'm not worried about technical disruption of bitcoin but social -- after all there still aren't many merchants accepting it.  BTC could just fade away... statistics showing a growing user base would convince merchants to offer it as a payment mechanism.  You are right, the inability to fully count/control the members is a great strength which is why I said "approximate numbers".  Hard numbers would be great, but if those are not available, it would still be useful to have the same kind of partly-fabricated numbers that businesses have used since the beginning of... well the beginning of VCs probably... to justify their business model.

It's not about 'my vision'.  Open-source is not an easy concept to grasp.  I'm sorry, but it's also not my fault.

From an implementation point of view and in the interest of balancing optimization against robustness, I might consider something like this:

 - Developers (in conjunction with other leaders) implement highly important decisions as code which simply honors settings which only secret-key holders can induce.

 - Once a decision has been reached ('the votes are counted') the next release might optimize the decision in code.

 - For flexibility and visibility, the code with deals with decisions like this would probably be implemented as modular plugins.

Methods for verifying integrity and absorbing 'updates' and the like are quite solid and old hat at this point.  If there is some focus on avoiding spaghetti code, and in particular, constructing a well thought out reference implementation for 'voting' which is auditable, that would go a long way toward developing the necessary level of confidence in the solution.


The 'network effect' (as I visualize it) is a highly valued moderating factor in a well functioning open-source solution.  Let's use Bitcoin as an example here:

Let's say that most people liked BIP17, but Gavin and relatively few others liked BIP16.  He could still implement it and the project would not come off it's rails because it although a majority is annoyed and some of their hopes for certain things are dashed, it's not a show-stopper and a large part of the community trusts him and his decisions allthesame.

Now instead lets say that Gavin made an obscure hack which transferred everyone's BTC to his address and pumped out an 'emergency fix' that everyone should update to ASAP.  Of course the code would fork almost immediately and a bad actor would be reduced from the community.  Probably in that bad of a scenerio transactions would be rolled back even.  Bitcoin would be damaged, but would survive.

---

As an aside, I personally build all of my releases from HEAD.  I do some cursory evaluation of the repository, but I've not the time or skill to identify well constructed exploits so my evaluations tend to just be to identify points of likely stability.

Part of the reason I do things this way is that I enjoy it, but another part is that if I stopped trusting the dev team, I would be prepared to look after the value I hold in Bitcoin anyway.  Or at least have a fighting chance of doing so.


I've already stated that all I have at this point is a neglected thought experiment, but you seem to keep asking for instrumented binaries (ok, and overstatement.)

I tossed off some implementation detail on-the-fly above.  Really though, if you cannot rather effortlessly envision at least the possibility of implementing 'veto' in a Bitcoin-like crypto-currency solution, and even a workable skeletal framework, it is somewhat pointless to try to enlighten you much further on this particular aspect of things.

Wow. Just when I thought it couldn't get any worse.

Why you gotta be trolling this highly enlightened conversation?

Brotha', that's how I roll.

Heh.  Someone (who can) seems to have taken a mulligan on this forum thread.  Unsurprising really.

I formally appologize fot this (and only this) thorny barb.  I had not realized that the conversation had been move and it seems both fair and appropriate to me.  More than fair, in fact, given that it was not moved to the 'alt' wasteland of filth which I would be hard pressed to argue as inappropriate:

  https://bitcointalk.org/index.php?topic=84941.0 (https://bitcointalk.org/index.php?topic=84941.0)

I've got one more relatively important think to say about 'bakcoin', and a lot of loose ends from the lively conversation with one of the sharper knives in the bitcointalk.org drawer which I may or may not fill in depending on what else I've got going.

A MultiBit user has experimented with using a SOCKS5 proxy and it appears to work (I haven't tried it personally).
She started MultiBit using:

java -jar -Dhttp.proxyHost=127.0.0.1 -Dhttp.proxyPort=8123 multibit-exe.jar

i.e. localhost and her SOCKS5 proxy was on port 8123.

The Mac and Linux installers have the multibit-exe.jar in. On Windows the jar is wrapped into an exe file so you could not use it directly as you cannot pass command line options on. (You could grab the multibit-exe.jar from one of the other installers and it would work).

The current version of MultiBit will leak a little as the help is taken directly from http://multibit.org.
At some point I will bundle the help in the installer so this leak will disappear.

If that scenario ever materialize (states shutting down, seizing or controlling mining pools and other full nodes), these full nodes could migrate to more friendly jurisdictions. If eventually these jurisdictions became rare or inexistent, they could migrate to the darknet.
Shutting down darknets is a much more complex thing to do politically speaking - only shameless dictatorships attempt to do it so far.

I think slush is already accepting connections to his pool via a Tor hidden service, isn't he? In his case it was not for hiding his location, since he has a public IP anyway, but it was a measure to avoid DDoS I think.

I like this idea.

I also like the idea of using Raspberry Pi or other cheap, disposable single-board computers with free, public Wi-Fi connections (library, airports, train stations, etc) as nodes. Of course, port forwading 8333 would be a challenge.