Bitcoin Forum

Check out my video, I discus this and more :)

http://youtu.be/WKIU5ha9F8Q (http://youtu.be/WKIU5ha9F8Q)

For those who don't have time to watch a video, can you post a summary or the transcript here?

I have no idea what your video is about.  I didn't bother to watch it.  I prefer to use the discussion forum for discussion.

As to your question, the answer is:

No.

My Diablo 3 folder is 14.9Go, is it too big ?
My Windows folder is 29.3Go, is it too big ?

Actually the video and the questions are okay, it's just that this topic has been discussed so much already that people are tired of this.
That's why you receive so much negative backslash on the video.

You dont need to download the Blockchain at all. Use Electrum client for example

That's not a very good advice in my opinion. Don't let third parties have the full knowledge of the chain.
Keeping it safe on your side , and checking it against other chains people keep on their side is the best way to make sure nobody can temper it.

If Blockchain and Electrum become the only actors that store the full chain, funny things could probably happen.

I understand why I received some negative comments on this video, but I do think those who don't read pages of forums would also like to know about some common discussions :)

U don't need to use Electrum at all. Use a credit card for example.

I don't think it's so much the size once it's on your computer but rather the time it takes to download from the network that can be the problem. If you don't have a fast internet connection it can take a long time.

People who have problems with the size can just use MultiBit:) It dos not have to sync like the BTC Qt dos.

Jason, first thanks for the all the videos you've made they were really helpful for me when I was studying btc at the beginning. To answer the question, I think it's fine as it is but I also think that future wallets will have default settings which won't download the full chain.

As long as it fits on 2TB hard drive, it is not too big. I think there are many of us, who value the bitcoin more than a hundred bucks (which is the price of such HDD nowadays). Blockchain bigger than that would be a bit cumbersome on current office PCs, I admit. But it would take 200 years at present pace to reach this limit.

Comparison of blockchain size to HDD capacity is irrelevant. Ppl who do that usualy don't understand what the issue is.
Anyway, blockchain size problem is unresolvable with currently used technologies/approaches. Core developers won't admit it, u waste time talking about the problem, better switch to other currency.

Or you can ignore people with glowing orange buttons who make unsupported claims.

C'mon. I got the orange button for fighting against the Bitcoin Foundation. And by the way, how can somebody prove that something does not exist? Care to prove the solution of the scalability problem DOES exist? Or will u just ignore this request like all other core devs did?

Most people use laptops with SSD, and cant afford to have a blockchain size of 40 GB or bigger.. that consumes 1/3 or 1/6 of their SSD size.
If we want a truly decentralized network, nobody should be excluded. The solution is therefore to make the blockchain dynamic and small instead of one gigantic mirror on every HDD in the network.

That's a very bold claim, that most people use laptops with SSD. I for one do not know any of them.
I agree that the thing to do is to make a lightweight version of the blockchain that would contain only current spendable transaction outputs. But anyone willing to host the full blockchain should have the means to do so.

Not at all, nowdays most people buy thin clients like iPad (16GB-32GB) or small laptops (120GB-240GB) instead of heavy desktop PCs with 2TB HDD.
Heavy desktops will become a minority among households, so if we care about decentralization we must consider what kind of clients most people will be using within a few years.
When people need terrabyte storage they buy USB-drives or NAS-drives - which are useless to host the bitcoin client.

i do like these video's but the summation of predicted size of the blockchain in 2016 is a little over the top.
i remember that as of spring 2012 i had to download 8GB of blockchain data. 18 months later it is 11GB so that is only 3GB increase this year...

I am not sure how you even came to the 1.3 Terrabytes total. to get to that figure each block would need to be a fully packed to the top, filled with transactions 25mb block size limit
25x6 =150MB each hour
144x 24 hours=3.6GB each day
3600x365 days=1.3TB each year

i think you are presuming the block limit for 2014-2015 will be lifted to 5mb and 2015-2016 will be lifted to 20mb and that is totally on the presumption that EVERY block has 100% utilisation of transactions filling up 100% of each block to require such a huge jump.

so here goes, a more accurate (yet still guessed) increase of the blockchain size per year

so far 2013: the maximum size of a block is 1MB, there are 6 blocks per hour (1 per 10 minute average) 24 hours in a day 365.. = 52,560GB MAXIMUM chain increase size per year if 100% utilised.
meaning in 2 years (2016) it will be under an 8th of a terrabyte(125GB) if each block was totally 100% filled to the 1mb per block size.

so far the 2009-2012 have shown small amount of transactions which dont utilise the full 1MB space per block, even in 2013 on average only a 5th-8th (12%-20%)of a 1MB block gets utilised meaning the total blockchain for 4 years of running is only 11gb, not hundreds of gigabytes.

so lets pretend transactions atleast doubled next year on average(lets say moved from the 12-20% upto 33% block fill) there maybe another 17GB added to the blockchain totalling 29GB for 2014-2015. and each block still has space per block.

again another year passes and transactions increase double fold equalling 35GB for the 2015-2016 or 64GB for the 2009-2015 total, again still room in the block for more transactions.

i do not think that it will be terrabytes of data, because somehow, somewhere you got the 5fold increase figure, and then used this against the presumption of 100% filled blocks.

now the knit pick is over. you do have potential as a presenter, but its still worth you double checking your figures and research before posting videos or your credibility will slowly decline

You don't have to use Bitcoin-qt you can use Electrum as local wallet

translation
dont own your own bitcoins use a third party service that can take the coins..

i personally would say for those people that get bitcoins as a wage and spend them on foodler.com and bitcoinstore, meaning the "consumer" class of people, then yes using third party services for weekly amounts is acceptable.

but for investors and long term savers i would say using the proper client with full blockchain and understanding the best practices of backing up and paper wallet funds long term is advantageous.

Saying electrum (or SPV client) is a third party service is misleading.
Please explain how electrum server would steal your coins.  The truth is you probably have no clue, you probable attack vector.

It doesn't matter if it's really big. You just have to buy another hard drive for your dedicated wallet.

13GB is not much to download. 100GB would be size I would reconsider using Electrum client instead.

???

What are you talking about?

In what way do you not own your own bitcoins with Electrum (or MultiBit)?

With both those wallets, you have exclusive control of your private keys.

If Electrum or MultiBit are "third party services", then so is Bitcoin-Qt.

it sounds like he doesn't know what electrum is, or how it works.

Sometimes ownership of private keys is not enough. If Electrum can fake received transactions in ur "wallet" they can cheat u. With Bitcoin-Qt u can be sure that if u see "6 confirmations" then there are indeed 6 confirmations.

Unless the attacker isolates you from the rest of the network.

If one is paranoid they could check tx against multiple electrum server or even run their own in the cloud to ensure they are getting good data.

Well the growth of the blockchain seems to be following an exponential trend, and if it continues like that it wont be surprising to see it larger than 1 TB by 2017.

This is exactly why I've put so much of my effort into developing a new type of mini-blockchain scheme (http://bitfreak.info/mbc-wiki/) which doesn't require every transaction ever made to be saved forever.

Just for the fun I started again bitcoinqt on a laptop i haven't used in a while.
It says 31525 blocks remaining , I wonder how long it will take.

With a max of 1mb per block the blockchain should not be able to grow faster than 52GB/year.

Although that is still very large.

That is true, but the max block size will eventually need to be increased if bitcoin is to support the growing level of transaction rates. So there are two options 1) cap the transaction rate and thus cap the blockchain growth rate or 2) increase the max block size and deal with an exponential growth in blockchain size. Neither option is ideal, the best way is to design the blockchain in such a way that it is dynamic and old blocks can be easily discarded from the chain but at the same time the account balances remain verifiable.

I completely agree with you.

1) make the chain dynamic like you say, but also...

2) we need to be able to distribute nodes. Even if we get the size down the real time transaction amount may become very large (Imagine several thousand tx/sec). In this case a distributed node where you connect into a node cluster depending on your resources. So an average computer that wants to participate can connect to a swarm of 10,000 other similar computers and they all become one node. In this manner us little guys can still verify blocks even they are incomming fast.

3) also I think we need to move away from giant blocks and to smaller faster blocks. The block reward can be split too (Although in the long run this does not matter as tx fees will pay for the network). So imagine that hundreds of 500kb blocks confirm every minute, and as those blocks get old (past depth:1,000) they start to get compressed and merged and eventually deleted (back to your dynamic blockchain idea)

I like the idea of super nodes or "clusters", but aren't mining pools essentially the same thing? And I agree with your last point about fast blocks but the problem is that if you make the block rate too fast it can cause a large number of orphaned blocks and other problems. Anything less than 1 minute seems to be a bit too fast imo, but I think the 10 minute block rate of bitcoin is a bit too slow. I think there is a natural balance between too fast and too slow, something close to 2 minutes is optimal imo.

I agree with you, several methods need to be implemented to deal with the blockchain monster.
The current method where all client and miners download all the blockchain is not sustainable.

I have a SSD in both my laptops, my brother has SSD's in his laptop, My boss uses SSD's in every laptop issued by our company. SSD's are everywhere.

All my friends at uni have SSD's in their laptops, my parents and girlfriends parents that have no idea about computers now have ssd's (Without us being an influence on their purchase) and even desktops are being sold with ssd as default these-days.

SSD are subject to Moore's law.  The blockchain is growing slower than Moore's law.   The "SSD issue" is not an issue.  When the blockchain is 1 TB your laptop will have a 20 TB SSD and it will only cost $159.99 (the SSD not the laptop).

Get rid of block size limit and blockchain will start growing faster than Moore's law. Mass adoption and 1 Mb limit are incompatible each to other. The best we can do is to find Equilibrium.

Still waiting for your follow-up on the other thread there, gollum. Thanks.

Are there even any full clients available for iPad?

For my case, I am happily running the full Bitcoin-Qt on my laptop. With the blockchain on its single internal SSD. I don't know what the heck your problem is.

SSD GiB/$ increases by a factor of 2x every eight months or so. Even if someone is too obstinate to not buy enough storage for their imagined needs today, Moore's postulate ensures that they will soon be able to. Probably not even be able to find a drive small enough to present a problem. Seen any 10 GiB drives for sale lately?

Not on any decade long scale.  In the short term we may grow faster than that BUT the current transaction volume is a tiny fraction of what current hardware can handle so there is a "cushion" built in.  We could increase tx volume by a factor of 20x and not be limited by hardware.  So if we grow at double (or triple) Moore's law for a couple years it simply eats into that cushion.  There is no scenario where Bitcoin growth can exceed Moore's law on a multi-decade timeline, there are only so many humans on the planet.

In 30, 40, 50 years, the discussion of not enough GBs look as silly as saying "video games will never take off because it in the future they will require thousands of floppy disks".  I bought a 1 GB drive for $300 in 1995.  That would be $300,000 per TB, and in less than 20 years the cost to store 1 TB has fallen to $50.  In 20, 30 years I fully expect the cost to store 1 PB to be <$300.  This isn't ancient history we are talking about just the effect of Moore's law in a single lifetime.

However in the long run most users will use SPV clients.  The security of the network doesn't require 100% of users running a full node, just that a critical mass continues to do so.  Between merchants, , developers, exchanges, service providers, hobbyists, and just die hards who refuse anything less than an equal peer on the network I don't see it being an issue.

OK. U r Satoshi, so u know that better than anyone else. :)

It's not that long,yet.

Im actually impressed that bitcoin has worked so well from start, and will work well for several years.
My point is that we must be proactive and change the protocol now before it becomes to big.

Remember that the userbase of popular trends grows exponentially - not linear. So when the userbase and transactions goes up to the sky the infrastructure must be prepared for that success.

or yet alone in this last 6 month period.

41 remaining , it took almost 1 day to synchronize after ~100 days offline.

Wanted to install bitcoin-qt with a computer with 8GB left. Realised it don't have that much space. Got my New computer and it have been syncronizing for a day now. Any idea how long will it take?

mine takes a couple of hours.  i use imdisk to create a 16gb ram drive and then run it off of that

Depends on your download speeds and the upload speeds of your peers.

exactly hard drive sizes will get bigger over time :)

I would say rather than the chain being too big, the CHAIN is TOO SMALL!!!

Transactions are taking too long!

Increase the minimum blocksize to 10mb and maximum to 1 gigabyte.

That seems rather silly if most blocks don't even fill 1 MB yet.  Perhaps it would be better if more mining pools would expand their current usage to fill more of the 1 MB block than they currently are.

10 most recent blocks:

• Block 270527 (https://blockchain.info/block-height/270527) = 412.5 KB
• Block 270528 (https://blockchain.info/block-height/270528) = 145.7 KB
• Block 270529 (https://blockchain.info/block-height/270529) = 211.6 KB
• Block 270530 (https://blockchain.info/block-height/270530) = 55.3 KB
• Block 270531 (https://blockchain.info/block-height/270531) = 243.3 KB
• Block 270532 (https://blockchain.info/block-height/270532) = 243.1 KB
• Block 270533 (https://blockchain.info/block-height/270533) = 243.2 KB
• Block 270534 (https://blockchain.info/block-height/270534) = 161.5 KB
• Block 270535 (https://blockchain.info/block-height/270535) = 13.4 KB
• Block 270536 (https://blockchain.info/block-height/270536) = 208.5 KB

Average block size: 193.8 KB

What sense is there in increasing the minimum blocksize to 10 MB if we aren't even using up 200 KB?

Future-proofing and to allow people to gain more confidence by allowing transactions to occur faster. At the moment, you can only allow for 360,000 transactions a day. That's going to be way small. If you want people to jump in, the developers better fix it or at least do away with the 1MB limit.

Paying extra transaction fees doesn't make anything go faster. It merely transfers your position on the waiting list, driving up transaction costs for everyone.

Remember, we the users of bitcoin are causing this problem. Why don't we just fix it by rolling out a new client.

Well it isn't future proofing or improving confidence to just raise a limit and then have the defacto size remain the same.  Lets see if pools can use the available resources first.

Not necessarily! If I hold a very large hdd then 20 gb isnt much!

a nas can be used to store and dwnld blockchain

Just expand storage using RAID...

This is another reason why we should have bigger block sizes as it will benefit the uptake of bitcoins. Everyone seems to be scared of huge blockchains, but I don't think it's anything to be scared of.

i'm not scared, just concerned how it will be handled in the future

It's not just the storage space , it's the download that might get problematic for a lot of users.

Yet all the pipes get bigger every year. Sure, we should make sure to monitor the situation. But so far, nothing about this seems broken to me.

Wayyy too big! Its taking forever!

If you ask me it could be too big for the average user in a few years, and the average user is what matters.
The moment it gets too big for "home" users, bitcoin can no longer be considered decentralized.

It depends on several factors of course but I think it needs a solution.
Altcoins could turn out to be a forced solution at some point if it isn't fixed.
 

On a long enough timeline, the average user will probably not run a full node.  That is just the reality.  A full node is an equal peer on a global payment network.  You can't have "global" and low requirements in the same sentence.  Most people don't really think about that aspect but a full node is a full and equal peer (no less or more important than any other peer) on what potentially could be the largest payment network in the history of the human race.   That is always going to have a cost.  Alternative solutions may reduce that cost to some degree but you can't have a massive network with millions of users and billions of transaction and expect the PEERS (as in PEER TO PEER) to have negligible requirements.

The good news is Satoshi envisioned a solution before the genesis block was even created.  Most users don't need to be full nodes, Bitcoin protocol supports a type of client called SPV (Simplified Payment Verification).  SPV nodes/clients download in realtime, only the parts of the blockchain that are related to their transactions.  They still validate information received from full nodes (the use of merkle trees and blockheaders allows SPV nodes to ensure the information they receive is accurate).  More importantly they keep their private keys secret, and sign their own transactions so they provide a higher level of security than eWallets.  The resource requirements are negligible compared to a full node.   So it comes down to a choice.  Do I want to be a user on the network at negligible cost or do I want to BE a part of the network as a full node with the costs that will require?  Many (probably most) users will opt for the former but Bitcoin doesn't need every user to be a full node, it only needs a critical mass.  IMHO there will always be that critical mass.  Enthusiasts, users with large amounts of Bitcoins, merchants, service providers, developers, and just "hardcore" believers will provide enough nodes for that critical mass.

Altcoins are not a solution.  They only have lower requirements because they have less users, less transaction volume, and less history.  If they ever reached Bitcoin scales the some requirements would apply.

I get that, but as full nodes disappear and big data stores is the only full nodes left, the model bitcoin uses is no longer the best one imo.
At that point an old-style server-client model would do a better job. And with that goes privacy, ability and need to manage you own account.

It could be that having several altcoins would solve this better. The question is how to easily transfer assets between them

That would be neat

Black and white thinking.  The world is full of shades of gray.  When the requirements become high enough that is has a real cost many users will not run full nodes but many still will.  It doesn't have to be everyone is a full node or the network is reduced to a half dozen nodes.  The reality is inbetween those extremes.


Well no.  SPV clients have methods of protecting privacy called bloom filters.


Having several altcoins wouldn't reduce the load.   If there are x tx on Bitcoin network and Bitcoin is replaced by a number of smaller altcoins, to have the same tx volume would mean the requirements to run all of them is the same.  To exchange coins between chains or mine all the chains would require the same amount of resources.  If you feel that amount of resources will reduce the network to a handful of nodes then the same thing would happen.   If all the mining and exchanging is done by a handful of nodes then you are in the same scenario.

The reality is that won't happen.  Higher resource requirements, and many users opting for the cheaper SPV option doesn't mean nobody will run full nodes.  IMHO there will always be that critical mass.  Enthusiasts, users with large amounts of Bitcoins, merchants, service providers, developers, and just "hardcore" believers will provide enough nodes for that critical mass of full nodes.

In other words right now the estimate on the number of full nodes in the Bitcoin network is probably on the order of tens of thousands of nodes.  Lets say 50,000 full nodes and lets imagine the number of users is 500,000.  That means 1 in 10 users is running a full node.  That ratio will drop if we jump to 1M users we probably won't see a mangnitude increase in number of full nodes, it might still be ~50,000.  The ratio of users to full nodes is now 100:1 but we still have 50,000 full nodes in 200 or so countries to provide decentralization.   Jump further in the future 10M users, 100M users, etc the ratio between users and full nodes is almost guaranteed to drop but that doesn't mean the number of full nodes will drop from 50,000 to 50.

It's easy to run a full node now. In the future, the "hardcore" believers might dedicate a hard drive and an internet connected computer just to run a full node.

You can regularly find torrents of 50+ gigabytes of complete seasons of old TV shows with at least a few thousand seeders, for comparison purposes.

This is correct. If I was an early adopter (or at least gotten a bit earlier than I did), I would have already set up a full node. The costs are pretty slim in comparison to what bitcoin has earned us (me in this case).

Torrents don't require to verify a lot of signatures.