CS 4970

Cryptocurrency

Mining

 

 

Aaron Bloomfield (aaron@virginia.edu)
@github | ↑ |

 

Mining difficulty

Disclaimer

  • All the numbers shown here were accurate and current as of the time this slide set was updated
    • February 12, 2025
    • But they may have changed since then
  • However, the RVN & BTC difficulty is from much earlier blocks
  • Otherwise, they are (generally) consistent throughout the slide set
    • And the exact numbers don’t matter as much to explain the concepts

Coins we’ll study

  • btc logo Bitcoin: we all know this
    • Hash algorithm: double SHA-256
  • rvn logo Ravencoin
    • A fork of Bitcoin, so works similarily; focuses on tokens
    • Designed to be ASIC resistant
    • Hash algorithm: kawpow (16 different rotating algorithms)
  • eth logo Ethereum
    • A generational improvement over Bitcoin
    • Allows for much more complicated programs
    • Hash algorithm: Keccak-256 (aka SHA3)
    • It just moved to proof-of-stake

Why a higher difficulty?

  • As more people mine – to get the reward – the time between blocks decreases
    • Each coin is designed to have a fixed time between blocks
    • Bitcoin: 10 minutes; Ravencoin: 1 minute; Ethereum: (previously) 12-15 seconds
  • Mining too fast will exhaust the available coin and and devalue the coin’s value

How to make it more difficult

  • Force the hash of the mined block to have a certain number of leading zeros
  • In Ravencoin block 3,714,773: 000000000000 9847addedabc03b320a9545175d969c7cf23dc9b862f301453cc
    • 12 leading hex zeros, 48 leading bits
  • In Bitcoin block 883,513: 00000000000000000000 d8dba3009cc023e03220d0ec6e5e3285e15ba0d9f0dd
    • 20 leading hex zeros, 80 leading bits

Random nonces

  • Consider a random nonce to generate a hash value
  • You would expect that half (50%) would have a leading 0 bit, and the other half would have a leading 1 bit
    • Expectation is that 1/4 (25%) have two leading 0 bits
    • Expectation is that 1/8 (12.5%) have three leading 0 bits
  • Expectation is that \(\frac{1}{2}^n\) have \(n\) leading 0 bits
  • Thus…
    • RVN example: \(\frac{1}{2^{48}}\), which is 1 in 281 trillion (\(2.81 \ast 10^{14}\)), would have 48 leading 0 bits
    • BTC example: \(\frac{1}{2^{80}}\), which is 1 in 1.2 septillion (\(1.2 \ast 10^{24}\)), would have 80 leading 0 bits

Bitcoin nonces

Mining: nbits

  • The nbits field is a 8-digit (32-bit) hex number, such as: 0x1b00a908
    • It’s like an enocded floating point number
  • The first 2 digits (1b) is the exponent; convert to decimal and subtract 3
    • \(0x1b -3 = 27-3=24\)
    • This means there will be 24 trailing 0x00 bytes (48 zero hex digits)
  • The last 6 digits of the nbits are the mantissa: 0x00a908
  • We left-fill with 0’s to 256 bits (64 hex digits)
    • With 48 trailing zero digits and 6 exponent digits, that leaves us 10 leading 0x0 hex digits (5 0x00 bytes)
  • This gets us: 0x000000000000a90800000000000000000 0000000000000000000000000000000

Mining: nbits & target

  • An nbits field of 0x1b00a908 thus gets us a target of 0x000000000000a908000000000000000 000000000000000000000000000000000
  • That target requires 12 leading zero hex digits (48 leading binary zero bits)
    • Chance to find: \(\approx 1/2^{48}\)
  • The first RVN target (RVN block 1) was 0x1d00ffff = 0x00000000ffff00000000000000000000 00000000000000000000000000000000
    • Chance to find: 1 in \(2^{32}\) or 1 in 4.3 billion
    • Note that some incorrectly report it as 0x1e00fff

Mining: targets & difficulty

Miner reported difficulty

  • Another way to look at it:
    • Let \(d_n\) be the network difficulty (say, 100k)
    • Let \(d_o\) be the original difficulty of block 1 (\(2^{32}=4.29G\))
  • The current target is as hard as the product of those two: \(t = d_n * d_o\)
    • For RVN block 1,982,364, the network difficulty was 100.3417k
      • (RVN historical difficulty can be found here)
    • Making the target as hard as \(100k * 4.3G = 4.29T\)

bdiff versus pdiff

  • The first RVN & BTC target was 0x1d00ffff = 0x00000000ffff0000000000000000000000 000000000000000000000000000000
    • This is the exact expansion of the nbits value
    • That yields a difficulty of 99,253.59955
    • This is bdiff (think: bitcoin-diff)
  • Traditionally it was represented as a target of 0x00000000ffffffffffffffffffffffffffffffffffffffffffffffffffffffff
    • This is a “non-truncated target”, and usually the values used by mining pools
    • That yields a difficulty of 99,255.11406
    • This is pdiff (think: pool-diff)

Mining pools

  • The chance of you getting a Bitcoin block is really low
  • But if you team up with other people, you can all contribute, and if the group gets a block the reward is distributed among the members
  • This is a mining pool
  • For Proof-of-Work (PoW) coins, members get a percentage of the value of the pooled profit based on the amount of work they put in
  • But how to compute that?
    • More on that in a bit…

Miner reported difficulty

  • A mining program will report the difficulty of a share
    • For now, a share is a nonce found that has a lot of leading 0’s, even if it is not enough to solve the block
    • “134.25 T”, for example
    • I’ll refer to this as the share difficulty, or \(d_s\), to avoid confusion
    • Share difficulty is set by the mining pool
  • If a share requires 42 leading zero bits, then it would report a share difficulty of \(2^{42} = 4.4\) T
    • 44 leading zero bits yields a share difficulty of \(17.6\) T
    • The target for RVN block 1,982,364 needed 48 leading zero bits
      • \(s_d=2^{48} = 281.5\) T

Mining shares

  • Any good nonce is reported to the mining pool, even if it is not enough to solve the block
    • Consider the 2miners RVN mining pool, which has pdiff (pool mining difficulty) of 4.295 G
    • So any nonce found that has a difficulty of \(4.295\) G$ = 4.295 * 10^9$ or higher will be reported to the mining pool
  • It is thus expected to take \(4.295G = 2^{32}\) hashes to get a share
    • So a 50 Mh/s GPU (the Nvidia 3090) would expect to get a share every 85.9 seconds
  • The share value keeps the bandwidth to the mining pool down while also allowing an estimate of the work performed

Environmental Issues

How much is mining killing the Earth?

  • Consider the Avalon A1466 miner
    • It uses 3,230 W and produces 150 Th/s
    • That’s 46.4 Gh/W, which is not super efficient any more
    • 3,230 W for 24 hours/day yields 77.5 kWh
    • At Charlottesville’s electric rates (15¢ per kWh) that’s $11.67 per day
  • This site estimates that the current world-wide BTC hash rate is 804 million Th/s
    • If it were only Avalon A1466 miners, then we need 5.37 million such miners, at 150 Th/s, to get 804 million Th/s
    • That’s 17.4 GW (\(10^{12}\)W), or 416 GWh per day, or about 152,000 GWh (152 TWh) per year

Death to the Environment!

  • The lower-bound estimate from the last slide is that BTC mining uses 17.4 GW (\(10^{12}\)W), or 416 GWh per day, or 152 TWh per year
  • That is more than 171 of the world’s 218 countries (source)
  • That is the equivalent to 2% of the total US electrical usage per year (source)

Death to the Environment!

  • These computations:
    • Assume that all the BTC mining is as efficient as the Avalon A1466 miner
      • To little efficiency and money is lost, not gained
    • Do not factor in the other mined coins
  • So perhaps the actual amount is twice our initial estimate?
    • That is the equivalent to 4% of the total US electrical usage per year (source)
    • That’s 166 TWh per year
    • But that’s a total guess, though!

Can mining not kill the environment?

  • Yes!
  • Use renewable resources, or pay carbon offsets
  • But many entities mine Bitcoin for the profit, and these things often lessen the profit
  • So few do this

Worldwide power consumption

  • In 2021, the world used 25,343 TWh of electricity (source)
    • The US used 3,979 TWh (15.7% of total)
    • Only China was higher at 7,806 TWh (30.8% of total)
  • We guessed that cryptocurrency mining uses 166 TWh per year
  • That is 0.63% of the worldwide total energy usage
    • 0.32% for Bitcoin only

Actual usage

Mining profitability

  • Bitcoin mining is profitable at scale
    • Not by the individual
  • Reasons:
    • Significant price discount when buying miners in bulk
    • Industry pays much less for electricity
    • Tariffs can be avoided easier
    • Heat can be re-used for other purposes

BTC mining in the US

Proof-of-\(X\)

Mining

  • In order for transactions to occur, there has to be a way for some entity to “certify” a block
    • So far the only certification method we’ve seen is proof-of-work mining
  • But how to decide who gets to do it?
    • Since there is a reward involved, everybody wants to be that one entity!

Proof of Work (PoW)

  • This is what we’ve seen so far:
    • Find a nonce (and transaction order) to get the hash less than the target
  • The successful miner has shown s/he has done the work…
    • … or just been lucky enough…
  • … to “complete” the block

Proof of Work (PoW) issues

  • High energy usage
    • Environmental issues as a result
    • It’s not a “green” solution by any means
  • Can cause bottlenecks when there is a lot of activity
    • Not (currently) a problem for Bitcoin
      • It was before witnesses were added, though
    • Bit it’s a problem for “things” that runs on top of Ethereum

“Useful” Proof-of-Work (PoW)

  • Use computing power for some “good”
  • xpm logo Example: Primecoin, launched in 2013
    • Purpose: generate primes for mathematical and scientific study
    • Whitepaper
    • Time between blocks: 1 minute
    • Difficulty change is for each block (not 2 weeks)
    • Generates two types of primes sequences: Cunningham chains and Bi-twin chains

Cunningham chain

  • Of the first kind: length \(n\) sequence \((p_1, \ldots, p_n)\) such that: \(p_{i+1} = 2p_i+1\)
    • Each prime is one more than twice the previous
    • Example: 2, 5, 11, 23, 47
  • Of the second kind: length \(n\) sequence \((p_1, \ldots, p_n)\) such that: \(p_{i+1} = 2p_i-1\)
    • Each prime is one less than twice the previous
    • Example: 19, 37, 73
  • Generalized: length \(n\) sequence \((p_1, \ldots, p_n)\) such that: \(p_{i+1} = a \ast p_i + b\)

Bi-twin chain

  • A sequence of primes of the form:
    • \(n-1,n+1,2n-1,2n+1,\ldots,2^kn-1,2^kn+1\)
    • Must have an even number of terms
  • All examples are really large numbers…

ASIC mining

  • Some cryptocurrencies use hash algorithms that special-purpose hardware can mine
    • ASIC = application-specific integrated circuit
    • Bitcoin, Ethereum (before the ETH2 upgrade)

ASIC-resistant mining

  • rvn logo Consider Ravencoin, which aims for ASIC-resistant mining
  • It uses 16 different hash algorithms
    • The last 8 digits of the previous block’s hash determine which 8 algorithms to use
    • That’s a LOT of hashes, and it is very power-intensive
  • Many are skeptical of it’s ASIC-resistance
    • If it were valuable enough, I’m sure somebody would develop an ASIC for it
  • The developers regularly update the algorithm to try to maintain ASIC-resistance
    • Original algorithm was x16r, then x16rv2, and now kawpow

Lite Hash Rate (LHR)

  • Nvidia makes graphics card GPUs
    • And everybody was buying their cards to mine cryptocurrency
      • Which priced out their primary market: gamers
  • Nvidia’s previous solution: create LHR cards
    • Make the card recognize mining of Ethash, and “lock” the GPU for 20 seconds
    • But this only affects Ethash, not any other mining algorithm!
  • Mining programs figured out how to dual-mine
    • It mines Ethash at half GPU usage and another coin for the other half
  • Nvidia gave up on this in May 2022 and removed LHR in recent drivers

Terminology

  • The act of certifying a block is called different terms with different certification mechanisms
  • Technically mining is certification on a proof-of-work blockchain
    • But, more generally, it encompasses any type of block certification
  • Other terms: signing, validating, etc.

Proof of Authority (PoA)

  • A small set (possibly singleton) of accounts are signers
    • They form a clique
    • One signer signs a block
    • Then a majority (>50%) of the clique must agree
  • Not great for a decentralized cryptocurrency!
  • Our course blockchain uses this
    • The clique was defined in the genesis.json file
    • It makes mining (aka signing) much easier in a development setting
  • Biggest uses: test networks, development environments, course blockchains

Proof of Stake (PoS)

  • Those with a stake in the cryptocurrency…
    • … meaning they hold enough of it…
  • Can be the ones to present a certified block to the network
    • They put their cryptocurrency “on the line” to vouch for the validity of the block
    • And they get the miner reward
  • This is called staking
  • But what if the block is invalid?
    • Then the certifier loses some of their cryptocurrency
      • Called “slashing”
    • While not unknown, it rarely happens
  • Those who can certify blocks are called validators

Proof of Stake (PoS)

  • The network chooses the validators based on:
    • Amount of currency held
    • Length of currency held
  • Implementations vary by cryptocurrency
    • Those with less “stake” will be validators less often
  • Coins that use this: Tezos, Cosmos, Peercoin, Tendermint, the ETH2 upgrade
    • ETH2 requires a minimum of 32 ETH to participate in PoS

eth logo Ethereum and PoS

  • Originally Ethereum was proof-of-work (hash algorithm: Ethash)
  • It had a difficulty bomb, aka “Ice Age”
    • At certain block numbers, the difficulty increases significantly
    • See this graph from the above link
  • Purpose: to force the developers to either:
    • Migrate it to a later date (as has happened so far)
    • Change to PoS (in the works for some time now)
  • The new Ethereum PoS proposal is called ETH2
    • Or the Paris Network Upgrade or the Casper protocol
    • It went live on September 15, 2022

eth logo Ethereum’s move to PoS

  • Phase 0: December 2020: create the “Beacon Chain”, a PoS blockchain
  • Phase 1: September 15, 2022: change Ethereum from PoW to PoS
  • Phase 2: sometime in 2023 2024 2025?: allow shards

ppc logo Peercoin

  • First implementation of PoS, from 2012
  • Consider coin age, measured in coin-days
    • 90 coins held for 10 days is 900 coin-days
    • Mining a block consumes the coin-days
      • Thus, those with a lower number of coins will eventually be able to mine a block

ppc logo Peercoin

  • Weakness: buy a lot of Peercoin, and wait a long time
    • You can then mine many successive blocks, using a portion of your coin age each time
    • Then execute a 51% attack (we’ll see this shortly)
  • Solution: checkpoints
    • Basically a periodic (few times per day) proof-of-work mined block
    • The nodes all agree on the “offical” blockchain, which then is immutable

Nothing-at-stake-problem

  • Consider a fork in the blockchain due to two simultaneously mined blocks
    • This will happen occasionally!
  • Which one chain should a validator mine for?
  • Ideal strategy: mine both
    • In PoS, it is not computationally demanding to mine both
    • This way you don’t mine the eventual “wrong” chain, and then not get the reward
  • With PoS, there is nothing at stake to prevent mining both chains
  • This can make double-spend attacks more feasible, as duplicate chains are not removed quickly

Nothing-at-stake problem

  • All coins must address this somehow
  • ETH2:
    • Part of your stake is a “deposit” on the validity of the block
    • If that block does not make it into the full block chain, you lose that deposit
      • Called slashing

Staking

  • Let’s say you wanted to be a validator for a PoS coin
  • If you had more of that coin…
    • … you could be “chosen” as the validator more often
    • … and thus get more of the validator reward
  • So you might “rent” coins from others to increase your balance

Staking

  • If you have some of a coin, you could “loan” it to others so they can better participate in PoS
    • This is also called staking
  • Advantages:
    • You get a payout: maybe 5% of your “loan” per year
  • Disadvantages:
    • You can’t get your “stake” back early
  • A good choice if you definitely are not going to sell in the near future

Staking Legality

  • On Feb 3, 2022, the IRS said it will not tax unsold, staked crypto (source)
    • You still pay when you sell it, of course
  • Such coins are “new property” not “income”
  • Once sold, presumably you would pay taxes on it similar to how you pay taxes on interest
    • But that was not mentioned in that article…

Proof-of-History

  • In sol logo Solana (SOL) the block verification combines PoS and proof of history (Solana’s whitepaper)
    • Each transaction has a hash of another transaction
      • This proves that the TXN happened after the “other” transaction
      • This forms a chain (graph, really) of events in time
      • The blocks (maybe TXNs?) also have an actual timestamp
  • This graph is:
    • Directed (any TXN points to a previous TXN)
    • Acyclic (you can’t have a loop in time)
  • Thus, the validator can always find the topological sort of a set of TXNs
    • Which means we can sequence them in the signed block

PoH and block verification

  • With btc logo Bitcoin, we have to compute the Merkle tree hash
    • Given \(n\) transactions, it takes \(\Theta(n)\) steps to build the tree
    • Given infinite computing power, it takes \(\Theta(\log n)\) steps
  • In sol logo Solana, each transaction can be verified in parallel
    • Just check that it goes one link back to another valid TXN
    • Given \(n\) transactions and an infinite computing power, it can be done in \(\Theta(1)\) time
  • A GPU can have 4,000 cores, and each can verify one TXN in parallel
    • Okay, not infinite, but still quite fast

Proof-of-Capacity

  • aka Proof-of-replication (PoRep), proof-of-storage, etc.
  • You allocate a certain amount of “capacity” to store the blockchain
    • Depending on amount, you are likely not holding all of the blockchain
  • But all parts of the blockchain are on many machines
    • Benefit: in the case of a failure, the “storage” – in the blockchain – is replicated
  • You pay to store data in the blockchain
  • storj logo Example: Storj

Proof-of-spacetime

  • Spacetime means allocating space over a period of time
    • And is not related to Einstein’s spacetime from relativity
  • Like proof-of-capacity, but…
    • Your reward is based on the amount of space dedicated times the amount of time
  • fil logo Example: Filecoin

neox logo Neoxa: “Proof-of-Game”

  • This is not a block certification method!
    • They just wanted to use the “proof-of-X” moniker
  • Neoxa’s whitepaper
  • Intent is to reward video game players for actions taken in the game itself
  • Pro: trying to get in early on the intersection of games and cryptocurrency
  • Cons: so many to list here, where to start?

“Green” coins

  • Generally, any coin that is not proof-of-work
  • Possibilities:
    • Proof-of-stake, like ETH2 and Peercoin
    • Proof-of-{capacity, replication, spacetime}, like Storj and Filecoin
    • Does not require proofs, such as xlm logo Stellar Lumens (XLM)
  • There were also some pro-environment coins
    • That gave rewards for, say, producing renewable energy
    • Difficult business model, so most have floundered

DAGs

Consider…

  • The first few blocks of an arbitrary cryptocurrency…
  • Here is the genesis block:

The first block mined

Block 0 is on the top, block 1 is on the bottom

Two blocks mined

We’ve just mined block 2

A transaction

Note that UTXOs are not only in the bottom level

Realism

  • Most (all?) cryptocurrencies are mined by the developers for the first many blocks
    • Ethereum’s first non-mining transaction was in block 46,147
    • Bitcoin only had a few hundred accounts in the first 50,000 (or so) blocks
  • But the previous images of the DAG still show the concepts of having many transactions

DAGs

  • The entire transaction list is a directed acyclic graph
    • aka DAG
  • In order to successfully mine, this has to be kept in memory
    • as that tells you if a UTXO has already been spent or not
  • Optimization: use hash tables (or similar)
    • Map a UTXO to the amount it makes available
    • And only keep track of the unspent UTXOs
  • So not really a DAG, then…

DAG size

  • You can see DAG sizes at https://minerstat.com/dag-size-calculator
  • Consider Ravencoin, whose DAG is currently at 4.87 Gb (as of Feb 2025)
    • Based on the expected DAG growth…
      • 5 Gb GPUS will no longer be able to mine it around May 2025
      • 6 Gb GPUS will no longer be able to mine it around Mar 2027
    • source

Attacks

Double spending

  • Definition: when one submits two TXNs spending the same UTXO
  • Prevented by:
    • The “valid block check” that miners & nodes perform checks for
      • Non-malicious nodes will not vote to approve a node that double spends
        • Or that spends an already-spent UTXO
    • In Bitcoin: having a witness hash for a transaction that cannot be changed

51% attack

  • This is an attack primarily on PoW blockchains
    • But still possible on PoS blockchains as well
  • Given the initial state in some block, say 1000

51% attack

  • Mallory spends a UTXO
    • On something that has a quick delivery, such as a digital asset
    • Or a in-person transaction where she takes possession of a physical item

51% attack

  • Seller waits some number of confirmations…
    • Say, 3 (in reality, likely many more)
  • Then the seller assumes the TXN will not change
    • Because there have been the desired number of confirmations
  • The digital asset is then sent
    • Or the physical item changes hands

51% attack

  • The original spender tries to spend the UTXO again
    • Mine it into a block
  • Node adoption policy: longest chain
    • The longest “chain” of blocks, starting from the genesis block, is the valid chain
  • But how to make the rest of the network accept the second spending of the UTXO?
    • Make that second TXN part of the new longest chain

51% attack

  • But how to make the rest of the network accept it?
    • Make that second TXN part of the new longest chain
  • Solution
    • Have to mine more than the rest of the network
    • Thus, must have 51% (or more) of the network mining capability, hence the name
  • If the attacker…
    • Who has 51% (or more) of the mining power…
  • … can ever create the longest chain…
  • Then they have “double spent” the UTXO!
    • Not true double spending, but close…

Longest Chain Adoption Policy

  • Miners are incentivized to always accept the longest chain as the “valid” chain
    • That chain has the greatest chance of becoming the “accepted” chain
      • After a certain number of confirmations
    • Any blocks mined on a different (and shorter) chain yield no reward, since their mined coins are not part of the longest chain
      • No reward means no incentive

51% attack

  • But is this viable?
  • But a mining pool might for other cryptocurrencies
  • Consider the rvn logo Ravencoin pool from 2miners
    • See what percentage they have now by dividing the pool hash rate by the network hash rate
    • On June 27th, 2021, they had over 51% (source)
      • Due to a purchase of hashpower from (likely) Nicehash
  • firo logo Firo had one in January 2021
  • btg logo Bitcoin Gold had one in May 2018 and January 2020

Required confirmations

  • Envision: we have two longest chains, and miners are split between the two
    • Neither has yet become “longer”
  • What if the miner for block 1001, version 1, were to immediately spend those mined coins
    • And then the version 2 chain were to become the longer (and thus accepted) chain?
  • Solution: wait until \(n\) confirmations have occurred before allowing spending of the BTC

Confirmations

  • Number of confirmations typically required (via 2miners):
    • xlm logo XLM: 0 blocks (5 seconds) (TXNs are irreversible)
    • btc logo BTC: 6 blocks (about 1 hour) for “standard” transactions
    • eth logo ETH: 240 blocks (about 1 hour) (maybe different w/PoS)
    • rvn logo RVN: 100 blocks (about 2 hours)
    • firo logo FIRO: 100 blocks (about 9 hours)
    • btc logo BTC: 60 blocks (about 10 hours) for “large” transactions
    • erg logo ERG: 720 blocks (about 1 day)
  • This is based on many factors: block speed, overall network hash rate, and the value of the coin

Forks & Orphan Nodes

Cryptocurrency “Forks”

  • There are four different meanings of “fork” with regard to cryptocurrencies
    1. Uncle / orphan / ommer blocks
    2. Soft fork
    3. Hard fork
    4. Source code fork

Type 1: Uncle / Orphan / Ommer

  • Given a starting block number (here, 1000)

Uncle / Orphan / Ommer blocks

  • Given a starting block number (here, 1000)
  • A miner mines a given transaction \(x\) into block 1001

Uncle / Orphan / Ommer blocks

  • Given a starting block number (here, 1000)
  • A miner mines a given transaction into a block
  • But another miner mines a different block without that TXN

Uncle / Orphan / Ommer blocks

  • Given a starting block number (here, 1000)
  • A miner mines a given transaction into a block
  • But another miner mines a different block without that TXN
  • That other block gets a confirmation

Uncle / Orphan / Ommer blocks

  • Given a starting block number (here, 1000)
  • A miner mines a given transaction into a block
  • But another miner mines a different block without that TXN
  • That other block gets a confirmation
  • And then one more

Uncle / Orphan / Ommer blocks

  • Given a starting block number (here, 1000)
  • A miner mines a given transaction into a block
  • But another miner mines a different block without that TXN
  • That other block gets a confirmation
  • And then one more
  • Eventually the block with the TXN is discarded
    • As it is not part of the longest chain anymore

Uncle / Orphan / Ommer blocks

  • Bitcoin calls these orphan blocks or orphan nodes
  • Ethereum used to call them uncle blocks or uncle nodes
    • But that is gender-specific
    • Now they are called ommer blocks or ommer nodes
      • Ommer is a gender-neutral version of aunt or uncle
        • But not a native English word
      • Like “sibling” instead of brother or sister
  • What happens to that TXN?
    • Other miners will likely pick it up
    • Wallet software (or equivalent) will re-send it if it “disappears”
      • Or at least notify you that it was not mined
  • This is an issue if we all could mine the blockchain in this class
    • There is no effective way to ensure it is not discarded

How to handle?

  • Bitcoin’s orphan blocks: once it is no longer part of the longest chain, is completely discarded
    • Reason: with 10 minute block time, this is much less likely to occur
    • Thus, no reward for the miner of the orphan block!
  • Ethereum’s ommer blocks, when PoW: not fully discarded
    • Miners still get a reward, albeit much smaller
    • The EVM keeps them around in memory
    • They are not the canonical “truth” of the blockchain, though
    • Reason: with a 15 second block time, this is much more likely to occur
    • They cannot happen on our proof-of-authority blockchain
  • Many others are based on Ethereum’s strategy

Type 2: Soft forks

  • Non-token crypto-currencies usually have a group of developers
    • Occasional changes are made to the protocol / mining algorithm / etc.
    • Bitcoin via BIPs, Ethereum via EIPs, etc.
  • Sometimes called a consensus fork
  • Generally, all the uses of the cryptocurrency (eventually) adopt the change
  • These forks are backward-compatible
    • At least until everybody switches over

Soft forks

  • How to get all the miners to adopt it?
    • Release the updated source code
    • The agreed-upon change will state that at some point, all further blocks must have that change
    • That point could be when:
      • Some number (say, 75%) of the nodes have adopted this change
      • After some date/time
      • ETH 2: when the Total Terminal Difficulty reaches (or exceeds) exactly \(5.875*10^{22}\)
  • Example: Bitcoin’s Segregated Witness from 2015

Type 3: Hard forks

- A change is proposed that some users do NOT want to adopt - Or that is not backwards-compatible, so some users choose one way and some the other way

eth logo Ethereum &
etc logo Ethereum Classic

  • June 2016: A well-known Ethereum organization had a programming bug that allowed millions to be drained from the account
    • This endangered the perception and trust of the Ethereum as a whole
    • We’ll see this error in detail in Blockchain Applications
  • The fix: roll back the blockchain to before the hack to fix it
  • This split the users; some didn’t want the fix
  • etc logo Ethereum Classic is the blockchain without the fix
  • eth logo Ethereum is the blockchain with the fix

btg logo Bitcoin Gold

  • Hard fork of Bitcoin on October 24, 2017 (at block height 491,407)
  • Reasons: change to an ASIC-resistant hash
  • It’s had a lot of growing pains:
    • Was hastily put together
    • Had a developer pre-mine, which always looks shady
    • Hit by 51% attack in May 2018; 388k BTG (about $18 million) was stolen; promptly de-listed from Bittrex
    • Hit by a 51% attack again in January 2020
    • In July 2020 an “emergency update” was released to counter yet another attack

Type 4: Source code fork

  • You can make a new cryptocurrency by:
    • Take a tested open source code base
    • Make a the modifications you want
  • The rvn logo Ravencoin (RVN) source code is based off of the Bitcoin source code; created in 2018
  • Changes:
    • 1000x as many coins
    • 1 minute block time instead of 10 minutes
    • Native support for assets
      • Could be NFTs, but can represent any digital or physical item
    • Different PoW hashes to be ASIC resistant

nmc logo Namecoin (2011)

  • Namecoin (NMC) was created in 2011, also as a fork of Bitcoin
  • Purpose: censorship-free domain DNS registration for the .bit top-level domain
    • Anybody can transact with Namecoin to reserve a domain in .bit
    • Fee: 0.01 NMC
    • Reservation lasts 36,000 blocks (about 200 days) unless updated or renewed
  • .bit domains are not widely supported
  • Namecoin not very popular these days
    • In 2015: 120k domains registered, but only 28 in use

https://github.com/aaronbloomfield/ccc