Author

Topic: Transaction fees vs block-chain size (Read 1218 times)

newbie
Activity: 58
Merit: 0
July 23, 2012, 05:24:31 AM
#8

Well no because that assumes there are no other transactions competing for space in the block.  Already today in order to get priority processing many users pay > 0.0005 BTC.  An attacker attempting to launch such an attack would simply find that other users raise fees as they see their confirmations slow down.  A smaller and smaller % of attackers fee are put into the block.  Worse many blocks would go over the 50% mark and the fees required for inclusion of low priority transactions would skyrocket.  The network is well protected against a spam attack.  If necessary the mandatory fees (which AND I CAN'T STRESS THIS ENOUGH only affect low priority txs) can be raised by a consensus of miners.  It has been lowered in the past.  Even without raising the fee the USD cost will rise in relation to the USD:BTC exchange rate.

The attacker would be just spamming, not trying to push out normal transactions from blocks. If he sends 1000 spam transactions per block at 0.0005 BTC/tx, then according to the described mechanism they will be added to the blocks so the blocks are filled up to 50% of the limit. If some block happens to be this size already, then none of the spam tx is included, however, it is still a big block.

Okay. Assuming that the miners don't like spam either, I agree it is unlikely that any big spam would go either unnoticed, or noticed and still packed into blocks by miners.
donator
Activity: 1218
Merit: 1079
Gerald Davis
July 22, 2012, 08:17:56 PM
#7
Aaah.
Well, I've just read the max block size is 1 MB (https://en.bitcoin.it/wiki/Weaknesses#Spamming_transactions).
If the attacker fills every block to 50% (500 kb) at that assumed price of 0.0005 BTC/tx, it's theoretically possible to have that +1 GB in about 14 days.
500,000 bytes/block * 2016 blocks/2weeks = 1 GB/2weeks
Anyway such gradation of fees within a block is a good idea, and thanks for the info.

Well no because that assumes there are no other transactions competing for space in the block.  Already today in order to get priority processing many users pay > 0.0005 BTC.  An attacker attempting to launch such an attack would simply find that other users raise fees as they see their confirmations slow down.  A smaller and smaller % of attackers fee are put into the block.  Worse many blocks would go over the 50% mark and the fees required for inclusion of low priority transactions would skyrocket.  The network is well protected against a spam attack.  If necessary the mandatory fees (which AND I CAN'T STRESS THIS ENOUGH only affect low priority txs) can be raised by a consensus of miners.  It has been lowered in the past.  Even without raising the fee the USD cost will rise in relation to the USD:BTC exchange rate.
newbie
Activity: 58
Merit: 0
July 22, 2012, 02:03:20 PM
#6
Aaah.
Well, I've just read the max block size is 1 MB (https://en.bitcoin.it/wiki/Weaknesses#Spamming_transactions).
If the attacker fills every block to 50% (500 kb) at that assumed price of 0.0005 BTC/tx, it's theoretically possible to have that +1 GB in about 14 days.
500,000 bytes/block * 2016 blocks/2weeks = 1 GB/2weeks
Anyway such gradation of fees within a block is a good idea, and thanks for the info.
donator
Activity: 1218
Merit: 1079
Gerald Davis
July 22, 2012, 01:35:11 PM
#5
No.

Your understanding of DOS prevention fees is incomplete.  When a block is more than 50% full the required fees (for low priority transactions) start to increase.  At 50% full the fee would be 0.0005 BTC.  At 80% full it would be 0.002 BTC. At 90% full it would be 0.05 BTC.  At 99% full it would be > 1 BTC.  You could spend millions and go bankrupt before you could spam the Bitcoin network to death.  Of course that ignores what competition for block space will do to fess on legit tx.  Users who want to ensure their tx goes first will increase their fees meaning even paying millions in fees may not be a guarantee that your spam tx are included in a block any time soon.
newbie
Activity: 58
Merit: 0
July 22, 2012, 12:29:40 PM
#4
Okay, it's a fair response.

I don't know what kinds of people run the 3000 full nodes - btw isn't it 15000? (http://bitcoinstatus.rowi[Suspicious link removed].uk/ http bitcoinstatus rowit co uk) Maybe it will be 3000 after it grows to 50 GB. Wink I think the full nodes are very important to the network, and their significant decrease, due to e.g. spammed block-chain, would be harmful. But it's just my opinion. Or maybe the users would quickly migrate to the light-weight clients, who will quickly attach to the remaining full nodes, who will quickly start charging for their services, and problem solved.

So I think that either the development of the light-weight client infrastructure is of a priority, or that tx fees raising should be at standby to quickly deter any big spammers.
donator
Activity: 2058
Merit: 1054
July 22, 2012, 11:54:54 AM
#3
No need to bump after just 2 hours. As it happens I was reading the thread before the new post.

Let's start with storage. If there are 3,000 nodes then the extra 1GB means 3TB additional aggregate storage. 3TB of storage cost about $200. So the would-be attacker caused $200 of damage with a $10000 investment. With output upkeep costs as I suggested elsewhere, the transactions won't even remain in the blockchain for very long. In short, blockchain storage isn't a problem at all.

Bandwidth and signature verification are a bit more costly, but those too are bearable with a proper setup. Right now a "proper setup" hasn't yet matured, but we're getting there.

Yes, in the future most people will not run full nodes. We may have a transaction propagation incentivization scheme as in the red balloons paper, and we may also have a market of node services. Also, organizations with a big stake in Bitcoin who do not want to take any chances will also run their own node with no direct profit.

And in any case, a more accurate fee structure will eventually be implemented, but there's no urgent problem for now.

It's a wholly interesting project!
It is, welcome to Bitcoin and the forum.
newbie
Activity: 58
Merit: 0
July 22, 2012, 11:43:53 AM
#2
No opinions? Is adoption of light-weight clients the answer? In my opinion not really, because then there is too little incentive to run a full node. I don't know who will volunteer to cram the harddisk with huge block-chain. Also miners need just fresh transactions, and honestly, if I were a miner I wouldn't trouble myself with packing transactions to earn some extra 0.1 BTC/block in fees.
newbie
Activity: 58
Merit: 0
July 22, 2012, 09:52:41 AM
#1
Hello bees, I have got no particular interest in Bitcoin, except from "omg wtf!!!" Roll Eyes It's a wholly interesting project!

Q: Is it possible to spam Bitcoin badly, or even to extinction?

Seeing that the minimum fee is 0.0005 BTC/tx, equal to about 0.005 USD/tx.
Evil Entity comes and dumps $ 10,000 to have a good fun.
10,000 USD / 0.005 USD/tx = 2,000,000 tx
2,000,000 tx * 500 byte/tx = 1,000,000,000 bytes
Not so much money for an extra 1 GB. So when would you stop using the software, at 5 GB, or 50 GB? ...

As I read somewhere, there is an imposed limit on the network of 7 tx/sec. It means just that the 1 GB spam would take some 4 days.

I think the transaction fees should reflect this, and be more normal levels - it's NOT free. Nobody is going to get hurt at 0.05 USD/tx, when your local bank asks dollars.
Just my 50 cents.
Jump to: