Bitcoin Forum

So the blockchain grows exponentially as more and more users are using it. And now its around 35 GB and it will grow faster and faster , but we also know that the hard disks will also be bigger and bigger, thus they will grow paralelly ,and with the implementation of supernodes, larger companies will store the blockchain for us, after it gets very large.

However, what about the download speed??  ???

Since most of us have thought about the size of the blockchain, but what about the download speed.


Think of it this way, the size of the blocks have to increase at some point to absorb more and more transactions, or the block creation interval has to become shorter.


If the block creation interval becomes shorter, then at some point we have to get to an interval of less than 1 second, and after that there will be more transactions / time, than the speed of the average user, thus he cant keep up with the exponentially increasing blockchain, and the network will fall apart. This is true with supernodes , because this doesnt depend on the bandwidth , but the connection speed, which is bounded by the speed of light.

OR

If the block max size will be increased, then it depends from the bandwidth, and we know that even the fastest internet is not that big perhaps 1000x largers than the average one. And the internet bandwidth speed cant be increased like the hard disk space.

Largest bandwitdh ever: 1.4 terrabyte /s
http://www.independent.co.uk/life-style/gadgets-and-tech/fastest-ever-commercial-internet-speeds-in-london-download-44-films-in-a-second-9077061.html

However you can easily buy 100 petabytes storage space if you got money, but the internet speed will never increase at such a great pace.


Summary
So either you decrease block creation speed, but then you hit the speed of light boundary, to send information, and the blockchain falls apart
OR
You increase block size, but then you will run out of bandwidth, that not even big companies could provide, and blockchain falls apart.

Is bitcoin doomed?  :'( :'(

Also after some point if the blockchain gets too big, nobody could download it ever again, because your download speed would be smaller than the size of 1 block.

If 1 block size will reach 1 GB, and you got a 1mb/s internet then in 10 minutes, you can only download 600 mb, and by the time you download 600mb another block gets formed. So you will be forever locked out of the blockchain and you could never download it.

Even if you put it in a torrent, nobody could download it because by the time you download it it doubles or triples in size

 ??? ???

While there is a problem with scalability in Bitcoin, the number of transactions required to create blocks that exceed the network speed is astronomically large. I ran some numbers to show this:

The USA handled 122.8 Billion non-cash transactions in 2012. I rounded this up to 150 Billion transactions.
In transactions per second: 4756 tx/s in 2013. Round up to 5000 tx/s
In transactions per block: 3000000
Thus, using the average transaction size of 0.2 Kilobytes, the block size would be 600000 Kb = 600 Mb.

Now, this is just the USA, with some inflated numbers, so the numbers worldwide are definitely higher, but not be too much.

Even with larger than 1 Gb blocks, the worldwide average network speed is 23.3 Mb/s, so I rounded down to 20 Mb/s.
The largest block size which could cause the scenario you outlined would be 12 Gb. Which could most definitely support all the worldwide non-cash transactions.

Of course, these numbers are averages, they could be higher, they could be lower. The internet speed depends on the location, the number of transactions depends on the time and place.

While the number transactions could pose a problem for block size and internet speeds, due to the astronomically large number of transactions to cause a problem, the core developers are not focusing on that yet. They will probably get around to addressing that issue, but they have other issues such as the size of the blockchain. This issue is currently being thought about and developers are attempting to resolve it. They will move onto other pressing issues first before coming to fixing the issue with download speeds.

If you have a solution to this issue you present, by all means, share it. Propose a BIP and submit it to the community and the core devs. It will eventually become implemented, and you will have just fixed a potentially big problem with Bitcoin for the future.

Ok that's 600 Mb/s for a average population of pool of 319 million, but we got 7 billion people, so on average that would be 22x bigger.

And let's ignore the fact that USA has more money to send / higher developed nation, i dont think with internet & bitcoin it would be that hard to get access to transactions. So we can safely include undeveloped regions like africa and south asia.

So in best case scenario we would face 13.2 Gb /block, if people would restrain themselves.

(Now i`m sure your statistic doesnt count HIGH FREQUENCY TRADING inside exchanges, nor other financial markets, only interbank transfers is it?)

But with bitcoin everything has to go through the blockchain, while you could have an exchange holding debt/ IOU /promissary notes, it will definitely increase it on the large run.

But what about malicious users?

Suppose a malicious user would buy 1000 Bitcoin solely for the purpose of sabotaging the network, he then would send those 1000 Bitcoins in 1 satoshi pieces (even though he would have to wait 3-4 days) or 1 satoshi with 10k satoshi fee, depending how it would be more efficient, back-and-forth between his addresses.

So he could generate 100,000,000,000 transactions in 1 round, then after 4 days another 100 billion.

Thus in 3 weeks he could generate 500 billion transactions ,that would make 1 block 2.4 gigabyte instantly.

In 2-3 months he could destroy Bitcoin literally.  :-[

My numbers do not account for these, but they typically occur off chain. These transactions typically happen within the exchanges own system. The numbers I use come from https://www.frbservices.org/files/communications/pdf/research/2013_payments_study_summary.pdf. These numbers include all credit card, debit card, ACH, checks, and prepaid cards from 2012.

Someone doing that would be banned almost immediately. Their node would receive a banscore from its peers, and would thus be disconnected and blocked from the entire network for at least 24 hours. Miners would might also drop his transactions. Dust transactions have a lower priority, and it is likely that the miners would not include the transactions sent by the attacker into their blocks.

Yea, but eventually they will have to load it into the blockchain, and the more exchanges pop up the bigger a block gets.



Yea but i`m not talking about ordinary hackers, but rather a 3 letter agency, or some bankers who would not like it.

They would just use Tor or proxies to send the money back and forth.Now i agree that it would need more resources, a few million $ perhaps to pull it off, but those 3 letter guys definitely have it, if they decide to sabotage it they will be prepared.

1 method could be to raise the mining fee or eliminate the 0 fee completely, but that would raise all costs then. For example if you would set the mining fee to 0.01 satoshi that would make a transaction cost the same as a bank transfer, so it would not be that cheap anymore, but atleast it would protect us from this disaster.

The other solution would be to keep the block size limit less than 1Gb. Even if the attacker tried to flood the network, the miners would drop their dust transaction unless they paid an exorbitant transaction fee. Furthermore, someone would notice 100 billion dust txs from the same address and inform the miners, who could just drop those txs from that address. The only way the attacker could succeed would be if they had 50% of the hashing power, in which case the network is already in trouble.

But who said he would use the same address and same IP address.

He would do it like this


A->B->C->D->E...

Each time send it to a newly generated address, and change IP addresses if or when necessary.

--------

If you hard cap the block size at 1 GB, then you have to lower the block creation interval at some point, because the malicious user would fill up the TX with 1 GB of orders, and more waiting in pending mode + the other people's transactions.

So the pending orders would exponentially increase, and sooner or later only the bad guy would execute transactions, and everyone else would have his TX in infinite pending mode.

But if you decrease block creation time, then eventually you hit a lower time interval = light speed, the time it would take buch of data to travel from 1 point of the world to another.

So it needs (with ultra future technology) 133 miliseconds to transmit data, but with malicious users hijacking the whole network, you would have to lower the blockc creation time until you hit 133 permanent limit. And after you hit that, we are literally fucked, because the transactions will only stack up, and pending orders will be infinitely pending.

The only long term viable option is to delete the old transactions that are older than 6 months. Sorry but a permanent blockchain with full record history is not viable in the long term.

The latency of the internet and Tor and the time it takes to establish a connection makes changing ip addresses quickly enough to broadcast that many transactions per second impossible. The attacker would not be able to connect to another proxy and the bitcoin network, and still be able to send several million transactions per second. He's would need to broadcast then all almost at the same time or several million per IP. Thus, the network would still block him.

That already is a hard cap on block size... at 1Mb.

The blockchain only stores transactions that are confirmed, e.g. in a block. All unconfirmed tensions remain in the mempool, and if not broadcasted repeatedly, will be removed and forgotten within a day or two. Again, since 1 satoshi transactions are low priority miners are less inclined to fill blocks with them. Miners can also choose to not include any of the attacker's txs. Thus the attacker needs to continually broadcast all of his transactions to get them in a block, and the continual spamming of millions of dust transactions will get him blocked.

We can't change block creation target rate of 10 mins without a major fight over this hard fork. Changing creation rate messes with block rewards miners receive. They will fight this  :D

The block size limit is now 1M. I believe the current bandwidth in a developed city network will have no problem accommodating 5M blocks. Remember the block size limit is only a "limit". Miners are interested in the 25 btc block reward and not fees. They do not have incentive to mine larger blocks for the fees because larger blocks are slow to propagate.

By the time block rewards are much lower and fees play a larger role, then we might see blocks approaching 5M. Hopefully better infrastructure would have been developed by then.

Multiple PCs , different locations?

I know its far fetched, but if a gov is seriously considering eliminating bitcoin, they do have alot of resources to do so.


Yea but again, multiple IP, different addresses, different PC's. With enough resources it can be done


So if 99% of the transactions are the attacker sending dust, then the mempool will be overloaded with other peoples transaction, that will never execute.

So either you have to increase that number or the miners increase the average fees.

The Bitcoin blockchain is not going to handle all transactions in the United States or the world with today's technology. Making hundreds of thousands of duplicates of every single transaction throughout the world is non-trivial. Even if Bitcoin blocks were massive we would run into the issue of transaction fees being too high to be reasonable, not to mention the outrageous bandwidth requirements that would be necessary.

That does not mean Bitcoin is dead or doomed (Why the FUD?). It means that other technologies must be used to facilitate the majority of "Bitcoin" transactions in the long run. This isn't ideal, but sometimes the perfect is the enemy of the good.

https://en.bitcoin.it/wiki/Off-Chain_Transactions
https://blog.coinbase.com/2013/08/06/you-can-now-send-micro-transactions-with-zero-fees/
https://lightning.network/

It's not FUD its a reasonable concern.

Ok even if 70% of users will prefer online wallet transactions (which could be handled internally, if a wallet can get a banking license/money transmitter or similar) i`m sure many people will just go with the "dumbed down, simple to use way" as not all of us are technical experts.

But as the services / products grow, eventually this one will become a concern.

Now it may need a few more years, up to 10-15 for this to become a serious issue, but its better to be prepared for everything.

I want bitcoin to succeed, and for that we must discuss & solve all threats that can jeopardize BITCOIN :)

Mike Hearn argues it'll be a "serious issue" by next year. Gerald Davis also made an impassioned post around here about how the Bitcoin block sizes need to be increased. Mircea Popescu argues they're both wrong. GMaxwell seems to have a neutral opinion, though he's argued repeatedly that decentralization of Bitcoin is paramount, but pointed out that actually hitting the limit won't cause any kind of devastation, like Mike Hearn argued.

I'd be willing to argue that the block size is actually already too big, or at least would like to see more discussion about the benefits of a more restricted block space. That the block limit needs to be raised is a foregone conclusion with poor support; Leaning way too heavily on FUD, appeals to emotion, and reliance on future undiscovered technology. All arising out of a vision for Bitcoin that competes with the likes of Visa on their terms, despite manifestly being unable to do so.

Greg Maxwell;
https://bitcointalk.org/index.php?topic=1054181.msg11318788#msg11318788
Gerald Davis;
https://bitcointalk.org/index.php?topic=946236.0
Mircea Popescu;
http://trilema.com/2015/gerald-davis-is-wrong-heres-why/
Mike Hearn;
https://medium.com/@octskyward/the-capacity-cliff-586d1bf7715e

I think the current bandwidth in most places will not have a problem. Mining is not widescale amongst the users, only a part of all bitcoin users is mining. Those who can afford the electricity and hardware. They probably can afford the higher bandwidth. While theoretically possible, it practically means the miners lose profit.

PoW is boring but it's also secure. PoS is not. What you're asking is like someone saying "AES encryption is boring, let's switch to ROT13"

Yea you took the words out of my mouth. Who cares if it's boring, bitcoin wasnt designed to entertain you, if you need entertainment go to a Disney park  :D :D

Now as the debate of 20 mb blocksize comes, bitcoin is getting centralized.

The wifi people with crappy connections below 1mb/s are now already unable to keep up with this blocksize.

INTERNET SPEED GROWS LINEARLY, BLOCKCHAIN SIZE GROWS EXPONENTIALLY, WHAT NOW?

The blockchain must grow linearly as long as there is a block size cap, so.. What are you trying to get at?

The cap will increase eventually, as more people use it. Offchain or not, the bitcoin network has to handle atleast 10000 transactions/minute, for it to become a worldwide used currency.

Just my thoughts. Otherwise it will become centralized, as only big guys could afford to host the network.

What's this "otherwise"?
We can have centralization of nodes, or centralization of transactions, take your pick.

Thats a trap question.

So you are saying either way centralization is inevitable? Its very sad.

I`d take the centralization of nodes anytime, vs the transactions, but that increases the risk significantly of rogue nodes.

The less nodes the less confirmation accuracy, which is the whole point of Bitcoin.

Looks like with the 7mb block size proposal, this theory can become closer to true i`m afraid.

Already the number of nodes are down, due to the size of the blockchain, soon it will be further down due to the size of individual blocks.

Bitcoin will inevitable become centralized, but if the price would go up suddenly then we can avoid centralization :)

I pick neither.

Clients could process transactions instead of miners which fully decentralises the operation of the network. Miners would still do as their name suggests, mine new bitcoins  ::). A block could essentially be one transaction but a few more might be more efficient. This would mean that the tx/sec is intrinsically linked to the number of nodes rather than hashing power and bandwidth load balanced amongst many clients.

You don't need the entire block chain to process or send/receive bitcoin. You just need traceability of the latest hashes. You could keep the block chain distributed amongst the network clients and reconstitute only those pieces required in a particular client to process transactions. If you were worried, you could audit the entire block chain at regular intervals by pulling every distributed block from the network and running the normal block verify. The block chain is essentially a hash table, it just needs to be distributed (dCache? DFSgc?). Several existing replicating storage systems (such as usenet) could be used to persist reference blocks (like the genisis block) to avoid encirclement of a client by naughty people :o.

well a 35 GB may take sometimes a bit like 1 day to resync intoa pen or cpu wallet but its ok the full 120 bitcoin years will have more GB and more internet speed too. and more solutions to come

I see that with blockchain pruning the size can be further compressed.

However now we face the blocksize raise, that is already a big problem for shitty pc's or smaller miners.


Now imagine if we were to get to the point where this "singularity" would happen. And the internet connection speed were the real issue.

There is only fiber optics you know, no other internet is faster than that.

Even though I believe this download speed will eventually cause a problem.

I am in support for the bigger block size (perhaps with some pruning techniques).


This will be a problem, but not in the foreseeable future, and until then there is no really excuse for not raising the block size.

My friend told me there is a way you can torrent the blockchain... I've also heard that torrenting is much faster...

But with the blockchain ever-changing, how can you torrent?

Doesnt matter if you torrent, it will eventually become the same problem, unless bandwidth will grow exponentially too, which it wont.

You can torrent the blockchain, and then the remaining blocks you download normally.

Torrent was for users who want their blockchain to be downloaded fast. It was applicable when headers-first synchronization was not implemented (before v0.10.0). After that, users were able to download from different peers and verify them simultaneously and torrenting was no longer needed. If you want to torrent it, it would be to do the full verification first then download the remaining blocks.

Since most of us have thought about the size of the blockchain, but what about the download speed.
Think of it this way, the size of the
blocks have to increase at some point to absorb more and more transactions , or the block creation interval has to become shorter .
If the block creation interval becomes shorter, then at some point we have to get to an interval of less than 1 second, and after that there will be more transactions / time, than the speed of the average user, thus he cant keep up with the exponentially increasing blockchain, and the network will fall apart. This is true with supernodes , because this doesnt depend on the bandwidth , but the connection speed, which is bounded by the speed of light.

Well after putting more thought into this I realized that this doesn't really matter  that much for sending data, or its insignificant problem.

However It can matter in terms of pinging peers. If the network is interpinging eachother,then if the bitcoin network becomes too big, the pings will really slow it down.

However, the internet is the parent network of BTC, so the internet itself would be in danger of this.


But my proposal is so theoretic that I dont think practically it will happen in the nearby 1000 years. Although it's still a good thing to wander upon.


Any final words? I will close the thread soon, so if you have any final words say it now.