Blockchain Security

PART 1

What is blockchain security

Some of you may have seen a number of my debates on twitter recently revolving around the security of blockchain ledgers. This is a very important discussion to have and is one of the key pillars that underpins the value proposition of a blockchain.

 

Security is a HUGE topic and I still have lots to learn. I expect this will be the first of a number of articles where I will explore this further.

 

Blockchain security is effectively a measure of the total sum of computational power and costs associated with a malicious actor attacking the chain. Attacking a blockchain comes in  number of flavors:

 

Reorganization attacks – commonly referred to as re-orgs, double-spends or 51% attacks for pure proof of work chains. This where an attacker acquires sufficient power to create a fake chain that allows them to spend the same coins (usually deposit to an exchange), sell the coins for another coin, withdraw those coins and then re-write the original transaction such that they may spend those coins again.

 

Denial of service or Spam Attacks – which are purely intended to bring down the blockchain service, similar to these attacks on websites and service provides online. This is a malicious attack that fills up block-space with fake transactions such that real users cannot get access, nodes go down and the ledger generally fails to perform its function.

 

Consensus Capture – Where a dominant security provider (mining pool, validator pool etc) has an overbearing ownership of the consensus mechanism and is able to initiate a hard fork with new consensus rules that are in the attackers best interest. 

 

All of these attack vectors are extremely difficult to defend against, let alone all at the same time. They require extreme decentralisation, intelligently designed incentive structures to defend via game theory and a high cost to successfully execute.

 

However achieving this security is, in my opinion, the single most important characteristic of a blockchain. This is how people can develop trust in the chain with their life savings and ensures the uncensorable nature of the ledger.

Who benefits from these attacks?

There are a few mechanisms for attackers to benefit from one or a combination of these attacks, I will start from most adverse to least for some of the common ones.

 

Nation state attacks – If China or the USA decided they were unhappy with the proposition of decentralised money, they may attempt to attack the blockchain probably via a DDoS or Consensus attack. This is a behemoth attack vector. It falls under very unlikely yet catastrophic if successful.

 

Competing chains fighting for credibility – We have seen this with Bitmain who supported BCash in the early days and launched a massive scale hash war against Bitcoin. This ultimately lost Bitmain their dominance in the mining industry and some hundreds of billions in revenue. BCash has similarly attacked BSV in this manner. For ASIC PoW chains, this attack is limited to coins with the same hashing algorithm which is SHA-256 for Bitcoin and its direct forks.

 

Short Interest Attacks – Where a very large leveraged short position is established and followed by some kind of attack which damages the chain credibility. The global impact of such an attack is likely to be a loss in price value where the short interest becomes wildly profitable due to the leverage. It is important to note that many small chains do not see attacks simply because there is no way to profit from a short position. As soon as margin trading is available for small chains, I strongly suspect these attacks will become more commonplace.

 

Reorganisations to profit from double spends which was covered earlier, literally spending coins twice.

PoW vs PoS vs Hybrid

In my opinion, there are three consensus mechanisms which are the most important to study regarding security which each have a dominant chain as the representative example.

 

PoW = Bitcoin (compared to Ravencoin and Ethereum 1.x)

PoS = Ethereum 2.0

Hybrid PoW / PoS = Decred and Dash

 

Proof-of-work = Bitcoin

Pure PoW chains get their security from miners providing electricity in the form of computational power. Bitcoin is the dominant pure PoW chain and it values security above all other elements, as shown by the 100% block reward dedicated to miners.

 

Proof of Work hashing on Bitcoin evolved from CPU à GPU à ASIC which represents an increase in efficiency of hardware. ASICs are generally extremely important to the game theory protection of a blockchain and I personally do not support chains that make themselves ASIC resistant. The reason for this is that miners must invest massive financial capital into ASICs and these are useful for solving the specific chain algorithm only. They are paperweights otherwise.

 

Therefore, a miner becomes the ultimate custodian of the chain. Not only are they protecting their current and future income (paid in the coin) but they are also protecting their CAPEX investment as well as the energy contracts they have signed for years ahead (OPEX).

 

In this way, ASIC miners are monogamous – they are married to a specific hash algorithm and will support whichever chain is most profitable using that algorithm, typically Bitcoin. We can see this in the hashrate dominance of BTC vs BCH and BSV. Miners on BTC could very easily dedicate a small fraction of their hash power to cripple BCH or BSV but as with game theory, they earn more by mining a small fraction on these chains to claim block rewards. When block rewards on BCH or BSV lose their appeal, these chains will die.

Figure 1 – BTC, BCH and BSV hashrates (source: https://twitter.com/yassineARK/status/1154520284629782528)

GPUs are able to compute many different hashing algorithms so are not locked to a single chain. For coins like Ravencoin and Ethereum which are GPU mined coins (ASIC resistant) they are susceptible to miners leaving he chain to mine whatever altcoin is the most profitable at the time.

 

In other words, GPU miners are polygamous, they have no reason to mine an unprofitable chain and no true allegiances. I can see long term, ASIC resistance will likely turn out to be a mistake by most chains that employ it. Note the Ethereum hashrate which has seen significant withdrawal of miners since September 2018.

Figure 2 – Ethereum hashrate has seen lots of miners leave the building since the 2018 bear.

 

Proof of Stake = Ethereum 2.0

Proof of stake is the latest and greatest craze in blockchain projects. In all honesty, it is for one reason…its super cheap and super easy to sell the idea of ‘passive income’. Ethereum 2.0 chain is the highest profile candidate to run this consensus mechanism and whilst still an idea is actively defended by the community.

 

PoW requires hardware, electricity and a hardcore bootstrapping phase where the chain is vulnerable to attacks by any GPU mining pool until it gets big enough. Proof of stake only requires an ICO to distribute coins, raise funds and fire up a bunch of validator nodes. Herein lies the biggest problem with PoS chains – the initial distribution.

 

It is extremely difficult to ‘fairly ‘distribute PoS coins to bootstrap the validator pool. Now validators provide a similar service to PoW miners where they create blocks and receive a reward in proportion to their relative stake on the network. The idea is that when you have capital at risk, you have skin-in-the-game and therefore will act in the best interest of the chain.

 

In theory, this should work as an attacker would need to own/purchase a large portion of the supply and thus risk their capital on an attack. Ethereum plans to implement a ‘slashing’ mechanism whereby bad actors who are discovered have their capital taken away from them as punishment.

 

Personally, I can see this working for the smaller attack vectors like DDoS, double spends and competing chain attacks. However, I remain unconvinced that a nation-state who wants a chain dead, could not deploy massive capital (perhaps even secretly printed money by central banks) to accumulate enough coins to destroy credibility and moneyness. (I’m exploring this in the wild here if interested https://twitter.com/_Checkmatey_/status/1155465218073485312).

 

One other attack vector I am not sure about is the perpetual wealth accumulation by PoS chains. It is very likely to become an oligarchy where those who have validator collateral will retain this in perpetuity. Many don’t like to admit this, but it is very much the same as PoW miner centralization where the wealthy become perpetually wealthier, have a growing share of the network security and ultimately, final say in governance (which will likely be to their benefit if history is any teacher).

 

PoS is a largely new consensus mechanism and its security will need to be trialed by fire if/when Ethereum 2.0 goes live.

 

Hybrid PoS and PoW

There are alternative consensus mechanisms which hybridize PoW and PoS for a new take on blockchain security such as Dash with master-nodes and Decred with tickets.

 

Both of these chains have unique security and value propositions and require further exploration. I will pick these up in a later newsletter (as usual, too much to get around in this industry before your mind explodes…stay tuned).

 

Closing thoughts

Blockchain security = trust and immutability. If you don’t need trust and immutability, you don’t need a blockchain. Therefore, by deduction, if you don’t have impenetrable security (from a probabilistic standpoint), your blockchain is dead in the water.

 

We have 10+ years of Bitcoins Proof of Work as our strongest point of reference for security. PoS is a hotly debated topic and I expect to see interesting developments as Ethereum research and development proceeds. Time and design will tell if they can land the catch and whether it is capable of withstanding the attacks Bitcoin has to date, because they will happen.

 

Hybrid consensus mechanisms are a whole new topic and I will look at Decred and Dash to explore their security mechanisms in Part 2.