Topic: Just a newbie question about blockchain size. (Read 2550 times)

I think this is relevant for the OP... http://en.wikipedia.org/wiki/Moore%27s_law

Just wanted to add this, because as it is already said, this is a non-issue, no impact whatsoever in the number of nodes or any impact at all in the network. If you can't/don't want to store the blockchain, use other clients.

you can use light version of bitcoin wallets like multibit,Electrum,Armory,Mycelium etc
they don't download big blockchain

I could have a new HDD with 1,000,000TB but full of files with no 29 GB left (if did exist).

Your HDD is small and old.  29 GB of HDD space is nothing today.
It equals to instalation of any new game.

not left, my HDD is full now

Many.... as in maybe 3% or less of home desktop users?  Shit, I only have 24GB of RAM.

As for the HDD space, yes, 29GB of HDD space is nothing.

Unless google fiber decides to come operate in this rural texas city I'm in, I'm relatively certain that bandwidth will always be my issue, not my PC specs.   768kbps upstream = detrimental to be running a public open node.  

You really don't have 29GB of HDD space?

anyway, I run a "full node" by just using connect= to one of my servers.  That way I don't connect to anyone else & don't have my upstream saturated (& slow people down in the process)

Yes.

It's called Electrum.

https://electrum.org/

---

MultiBit and Electrum are two options you can look into.

How to have a wallet without run a full node? do you suggest some way?

Satoshi has answered this newbie question in 2008:

https://bitcoin.org/bitcoin.pdf (Section 8: Simplified Payment Verification)



Many people already have 29GB of RAM in their PC, 29GB of HD is just nothing. If you can't even spare 29GB to hold the blockchain, you shouldn't run a full node.

I`ve already mentioned a solution to this somewhere.

We need to have a offline private key storage with offline signatures like Armory delivers.

However the broadcasting feature should be done by big storage companies who can afford to store the blockchain.

Using services like this:
https://blockchain.info/pushtx

This feature needs to be integrated into a lightweight wallet, so that once you securely sign your transaction offline, you can broadcast it on your online PC, from a lightweight wallet, without the need to have the blockchain on your PC.

Is there any lightweight wallet that offers this feature? Or is there any plan to do so in the future?

I'd reccomend you to look into multibit , online wallets are NOT safe.

Maybe "datacenter" was a exageration but the actual blockchain is already big to me. In theory we wont have this problem for 114 years but my pc is already with lacking space and I will not buy other HD, I do prefer to use some on-line wallet or paper wallet or anything else, still if there are some rate for use.
So less one with blockchain.
My phone can not store the blockchain too.

I believe the 7 day average block size is currently approximately 377.1 kilobytes.

As recently as February 8, the 7 day average block size was 293.1 kilobytes, and the largest 7 day average was 445.7 kilobytes back on January 27.

This would seem to imply that the usage of bitcoin could easily double from what it has been in 2015, and the blocks would still be less than 1 megabyte on average.  If the usage increased significantly above double what it has been, and the block size in not increased, then transaction fees would begin to increase a bit.  The amount of spam an non-financial activity in the blockchain would significantly drop.

If usage increased much beyond 265 transactions per minute (4.4 transactions per second*) and the maximum block size is not increased, then it would become difficult (impossible?) for the average user to use bitcoin for simple purchases and other financial transactions.

Of course, I've already mentioned that:


EDIT:

*This 4.4 transactions per second number was calculated as follows:

• Assume that the average user is using compressed key addresses
• Assume that (with effort and financial incentive) it is possible for most transactions to consist of 2 inputs and 2 outputs

A block has an 80 byte header.  This leaves 999920 bytes for transactions.
An input that was received at a compressed key address requires at most 149 bytes to spend
A standard address output requires 34 bytes.
A transaction has 10 bytes of additional (non-input or output) information

2 inputs at 149 bytes each plus 2 outputs at 34 bytes each plus 10 bytes of transaction overhead results in a transaction of 376 bytes.

999920 bytes of transaction space divided by 376 bytes per transaction results in room for 2659.36 transactions per block.  Since blocks occur on average every 10 minutes that's 265.936 transactions every minute.  Divide that by 60 seconds per minute and the result is 4.4 transactions per second.


EDIT 2: While putting together the above edit, I discovered an error in my math.  I forgot to divide by the 10 minutes per block when calculating the number in the parenthesis and accidentally ended up with 44 transactions per second.  The number has been corrected to 4.4 transactions per second.

Which limits the amount of transactions to few thousand per block (and per 10 minutes), according to blockchain.info. Which is ok for marginally used network.

The blockchain is currently only 29 GB in size.  According to Western Digital, you can purchase a 6 TB internal hard drive for a PC for only $270 today.  That doesn't sound like "datacenter infrastructure" to me.

Currently the blocksize is limited to a maximum of 1 MB per block, and the protocol limits the number of blocks to approximately 1 block every 10 minutes. Doing some simple arithmetic:


It appears that unless/until the protocol rules change, the blockchain can't grow by much more than 52.56 gigabytes per year.

Since 6 TB is 6,000 gigabytes, we can see with some simple arithmetic:


That means that a single $270 hard drive purchased today would be able to store the entire blockchain for the next 114 years.

How much has the size of a hard drive grown in the past 114 years, and how much has the cost per byte decreased over that same amount of time?  Taking that into consideration, how big do you suppose non-volatile storage (can we even reliably predict if hard drives will be the future storage media) will be 114 years from now?  How much do you suppose that storage will cost per byte?

Also, keep in mind that most blocks are not even using the entire 1 megabyte available to them yet.  Until they are, this further increases the amount of time until the 6 TB hard drive is full.

There has been some recent talk about increasing the maximum block size.  If this occurs, blocks might eventually grow bigger than 1 megabyte.  It is difficult to predict how quickly they will grow or how large they will grow (it depends on how quickly bitcoin becomes popular and how popular it becomes).  Obviously if blocks increase above 1 megabyte, it will reduce the amount of time until the blockchain fills a 6 TB hard drive.

It is also important to remember that bitcoin was designed with the intention that full nodes would only be run in large data centers.  Satoshi expected that there would eventually be less than 100,000 full nodes on the bitcoin network:

The blockchain is becoming bigger and the trend is that less people store it in its own computer and use online wallets or some other alternatives (mainly in cell phones which have little memory), possibly in the future will be necessary datacenter infrastructure to store the entire blockchain.

As far as I know the bitcoin network need many nodes having the entire blockchain to be reliable, so the huge blockchain will affect the bitcoin network?