/giftcoinmining is a subreddit dedicated to mining the Giftcoin crypto-currency.
1GB blocks would be able to confirm more than 5000tx/s. That would be VISA-level scale
(which handles, on average, 1736tx/s). We often hear that we shouldn't raise the blocksize because then nodes would become too expensive to run. But how expensive exactly?
We have the following costs to take into account:
For now, I'm going to assume a non-pruned
full node (i.e. a node that stores all transactions
of the blockchain) for personal use, i.e. for a computer built at home. I'll add in the calculations for a pruned node at the end, which would likely be the prefered option for people who merely want to verify the blockchain for themselves. If you don't care about the assumptions and calculations, you can just jump right to the end of this post. If you spotted any error, please inform me and I'll update my calculation.
There's, on average, one block every 10 minutes, that is 144 every day and 4320 blocks every thirty days. I was able to find a 3TB HDD for $47,50 on Amazon
, that is $0.018/GB. Storing all blocks with all transactions of a month (4320GB) would be $78.96/mo
. Prices for storage halved from 2014 to 2017
, so we can assume that to half in 2022, thus we can reasonably assume it'd cost around $40/mo.
But would such an inexpensive hard disk be able to keep up with writing all the data? I found a comparable cheap HDD which can write 127MB/s sequentially
(which would be the writing mode of Bitcoin). That would be enough even for 76GB blocks!
Edit: For the UTXO set, we need very fast storage for both reading and writing. Peter__R
, in his comment below,
estimates this to be 1TB for 4 billion users (which would make ~46,000tx/s if everyone would make 1tx/day, so id'd require about 10GB blocks). jtoomim
seems more pessimistic
on that front, he says that much of that has to be in RAM. I'll add the $315 I've calculated below to account for that (which would be rather optimistic, keep in mind).
Bandwidth is more complicated, because that can't just be shipped around like HDDs. I'll just take prices for my country, Germany, using the provider T-online, because I don't know how it works in the US. You can plug in your own numbers based on the calculations below.
1GB blocks/10 minute mean 1.7MB/s
. However, this is an average, and we need some wiggle room for transaction spikes, for example at Christmas or Black Friday. VISA handles 150 million transactions per day, that is 1736tx/s, but can handle up to 24,000tx/s (source
). So we should be able to handle 13.8x
the average throughput, which would be 1.7MB/s x 13.8 = 23.46M/s
, or 187.68Mbit/s. The plan on T-online for 250Mbit/s
) would be 54.95€/mo (plus setup minus a discount for the first 6 months which seems to cancel out so we'll ignore it), which would be $61.78/mo
. This plan is an actual flatrate
, so we don't have to worry about hitting any download limit.
Note, however, that we don't order bandwidth for only our Bitcoin node, but also for personal use. If we only needed 2MB/s for personal use, the plan would be 34.95€, thus our node would actually only cost the difference of 20€ per month, or $22.50/mo
. Nielsen's Law of Internet Bandwidth
claims that a high-end user's connection speed grows by 50% per year
. If we assume this is true for pricing too, the bandwidth cost for ~200Mbit/s/mo. would go down to
12.5% (forgot how exponential growth works) 29.6%
of its today's cost by 2022, which decreases our number to $
point out that the node would have a much higher bandwidth for announcing transactions and uploading historical blocks. In theory, it would not be necessary to do any of those things and still be able to verify one's own transactions, by never broadcasting any transactions. That would be quite leechy behaviour, though. If we were to pick a higher data plan to get 1000MBit/s downstream and 500MBit/s upstream, it would cost 119.95€/mo.
, however this plan isn't widely available
yet (both links in German). 500MBit/s of upstream would give us max. 21 connected nodes at transaction spikes, or max. 294 connected nodes at average load. That would cost $39.85
in 2022 (with correct exponential growth).
CPU/Memory will be bought once and can then run for tens of years
, so we'll count these as setup costs. The specs needed, of course, depend on the optimization of the node software, but we'll assume the current bottlenecks will have been removed once running a node actually becomes demanding hardware-wise. This paper
establishes that a 2.4GHz Intel Westmere (Xeon E5620) CPU can verify 71000 signatures per second... which can be bought for $32.88 a pair
on Ebay (note: this CPU is from Q1'10
). We'd need to verify 76659tx/s at spikes (taking the 13.8x number), so that pair of CPUs (handle 142,000tx/s) seem to just fit right in (given one signature per tx). We'd also have to account for multiple signatures per transaction and all the other parts of verification of transactions, but it seems like the CPU costs are neglegible anyway if we don't buy the freshest hardware available. ~$100
at current prices seem reasonable. Given Moore's Law, we can assume that prices for CPUs half every two years (transistor count x1.4162
), so in three years, the CPU(s) should cost around $35.22
For memory, we again have to take into account the transaction spikes. If we're very unlucky, and transactions spike and
there won't be a block for ~1h, the mempool can become very large. If we take the factor of 13.8x from above, and 1h of unconfirmed transactions (20,000,000tx usually, 276,000,000tx on spikes), we'd need 82.8GB (for 300B per transaction).
I found 32GB of RAM (with ECC) for $106
, so three of those give us 96GB of RAM for $318
and plenty remaining space for building hash trees, connection management and the operating system. Buying used hardware doesn't seem to decrease the cost significantly (we actually do need a lot of RAM, compared to CPU power).
Price of RAM seems to decrease by a factor of x100 every 10 years
), so we can expect 96GB to cost around $79.89
) in 2022.
Of course, CPU and memory need to be compatible, which I haven't taken into account. Chug a mainboard (~$150) and a power supply (~$50) into the mix, and the total would be just over $600
for today's prices. Even if mainboard and power supply prices remain the same, we'd still only have to pay around $315
for the whole setup in 2022.
I found the following power consumptions:
So we'd have
129W 147.6W + N*6W
. Electricity cost average at 12ct/kWh
in the US, in Germany this is higher at 30.22ct/kWh
. In the US, it would cost
$11.14 $12.75 + N*$0.52
(P*12ct/kWh / 1000 * 24h/day *30days / 100ct/$), in Germany
28.06€ 32.11€ + N*1.30€
At the end of the first year, it would cost
in the US and
At the end of the second year, it would cost
. for the US and
for Germany. It increases by $8.98/mo. per year in the US and by 22.46€/mo. per year in Germany.
Electricity prices in Germany have increased over time due to increased taxation; in the US the price increase has been below inflation rate
the last two decades. As it's difficult to predict price changes here, I'm going to assume prices will remain the same.
In summary, we get:
- Storage: $78.96/mo., $40/mo in 2022, (E:) +$315 initially for NVMe SSDs
- Bandwidth: $22.50/mo.,
$2.81/mo. $6.66/mo. in 2022, Edit: or $95.37/mo. for additional broadcasting, or $28.25/mo. in 2022 prices.
$20.12/mo. (1st year, US), $29.11/mo (2nd year, US); 50.52€/mo. (1st year, DE), 72.98€/mo (2nd year; DE) $21.73/mo. (1st year, US), $30.72/mo. (2nd year, US); 54.57€/mo. (1st year, DE), 77.03€/mo. (2nd year, DE).
- CPU: Initially $600, $315 in 2022
If we add everything up, for today's prices, we get (E: updated all following numbers, but only changed slightly
) $132/mo. (US)
, $187/mo. (DE)
for the second year and
. $78/mo. (US)
$115.79/mo. $124/mo. (DE)
It definitely is quite a bit of money, but consider what that machine would actually do; it would basically do the equivalent of VISA's payment verification multiple times over
, which is an amazing feat. Also, piano lessons cost around $50-$100 each, so if we consider a Bitcoin hobbyist, he would still pay much less for his hobby than a piano player, who'd pay about $400 per month. So it's entirely reasonable to assume that even if we had 1GB blocks, there would still be lots of people running full-nodes just so.
How about pruned nodes? Here, we only have to store the Unspent Transaction Output Set
(UTXO set), which currently clocks in at 2.8GB
. If blocks get 1000 times bigger, we can assume the UTXO set to become 2.8TB. I'll assume ordinary HDD's aren't goint to cut it for reading/writing the UTXO set at that scale, so we'll take some NVMe SSDs
for that, currently priced at $105/TB.
Three of them would increase our setup by $315 to $915 , but decrease our monthly costs.
E: However this UTXO set is also required for the non-pruned node, therefore the setup costs stay at $915
. Even in the highest power state, the 3 SSDs will need only 18.6W
in total, so we'll get a constant 147.6W
for the whole system.
In total, this is:
- New Storage: $0/mo.
- Bandwidth: $22.50/mo., (E:) $6.66/mo. in 2022, Edit: or $95.37/mo. for additional broadcasting, or $28.25/mo. in 2022 prices. (same as above)
- Electricity: $12.75/mo. (US), 32.11€/mo. (DE)
- CPU: Initially $915
In total, this is $35.25/mo.
in the US and $58.57/mo.
in Germany for today's prices, or (E:
(US) and (E:
(DE) in 2022's prices. Which looks very affordable even for a non-hobbyist.
E: spelling E²: I've added the 3 NVEe SSDs for the UTXO set, as pointed out by others and fixed an error with exponentials, as I figured out.
Accurate Bitcoin mining calculator trusted by millions of cryptocurrency miners since May 2013 - developed by an OG Bitcoin miner looking to maximize on mining profits and calculate ROI for new ASIC miners. Updated in 2020, the newest version of the Bitcoin mining calculator makes it simple and easy to quickly calculate mining profitability for your Bitcoin mining hardware. Calculate your Bitcoin loan with this easy to use calculator. Loan calculations form $2,000, using crypto as collateral, without needing to sell any. Bitcoin Average block time (minutes) Chart. Transactions Block Size Sent from addresses Difficulty Hashrate Price in USD Mining Profitability Sent in USD Avg. Transaction Fee Median Transaction Fee Block Time Market Capitalization Avg. Transaction Value Median Transaction Value Tweets GTrends Active Addresses Top100ToTotal Fee in Reward TradeBlock is the world's leading provider of institutional trading tools for digital currencies. Find out what your expected return is depending on your hash rate and electricity cost. Find out if it's profitable to mine Bitcoin, Ethereum, Litecoin, DASH or Monero. Do you think you've got what it takes to join the tough world of cryptocurrency mining?
This is part 3 of the Blockchain tutorial explaining what hashes are. In this video series different topics will be explained which will help you to understa... What it really takes to mine a Bitcoin in 10 Minutes. Firstly I'll show you a special free method to mine Bitcoin and send funds directly to your wallet in 1... Bitcoin Mining profitability Calculator http://www.vnbitcoin.org/bitcoincalculator.php Exchange bitcoin for usd http://www.vnbitcoin.org/sellbitcoinforusd.ph... #bitcoin calculator #bitcoin converter #bitcoin price history #how much is a bitcoin #bitcoin current value #btc value #btc price usd #bitcoin worth #bitcoin stock price #btc to usd converter #buy ... bitcoin block, bitcoin bot, bitcoin betting, bitcoin btcx, bitcoin blender, bitcoin blockchain size, b bitcoin, bitcoin b font, plan b bitcoin, plan b bitcoin podcast, bitcoin miner .b ...