Layer 2: The Blockchain Network
Blockchain networks are rapidly becoming ubiquitous, and have been applied to a vast array of problems and a varied cross-section of industries. The very first blockchain was used in the creation of Bitcoin, and continues to power the leading cryptocurrency to this day. This inaugural blockchain employs the Proof-of-Work (PoW) consensus protocol, an invention that solved a long outstanding problem in computer science and ensures continued security through ongoing resource allocation.
Since the early successes of Bitcoin's blockchain, attempts to find better consensus protocols have spurred a renaissance of innovation. While obviously vital to the growth and development of the industry, we feel that attempts to create such consensus protocols for the purposes of leaving PoW behind are, while innovative, somewhat misguided. Instead of re-inventing the wheel (namely, Proof-of-Work consensus) what is needed is a way to use the wheel in a more effective manner -- to harness it to build a machine that changes the world, and that connects the world.
At Zenotta we believe that PoW is, and always will be, the fundamental consensus philosophy that gives a blockchain network its power. But just as in a responsible, functioning, fair society we do not allow runaway greed & monopolistic endeavours to dominate entirely, so too should the PoW consensus mechanism be tempered, to reduce monopolistic tendencies, prevent mining re-centralization, and avert adverse environmental effects (more on this in the next section below). However, the first question that must be addressed is that of scalability.
In dealing with data, rather than tokens, a blockchain network must be able to scale to many orders of magnitude higher in transactions per second than current Proof-of-Work setups. Data transactions would have a far higher velocity than monetary transactions, and traditional Proof-of-Work consensus approaches were designed to be deliberately slow, in order to give miners time to act in the event of a malicious party attempting to subvert the blockchain, and to provide a relatively equal opportunity for people to mine on standard spec. CPUs. In order to achieve this high scaling, we begin with the important insight that of the three roles carried out by a blockchain; namely, block create (package transactions), block write (store on the ledger), and block verify (mine), it is only the latter that needs to be distributed in order to solve the Byzantine Generals' Problem and achieve the required trustless security.
It is therefore possible to separate out these three roles in a way that vastly simplifies and speeds up the mining throughput. Transactions per second can conservatively reach into the tens of thousands, while the Byzantine security of the ledger is preserved. Specialised nodes fulfilling each of the roles can perform their tasks in a dedicated fashion and communicate in parallel where possible, achieving optimal load-balancing and efficient time utilization. The Zenotta blockchain employs such an approach, as outlined below.
The Zenotta network protocol consists of three sub-networks built up from specialised nodes:
- 1.Compute Nodes that handle block creation, through packaging transactions
- 2.Mining Nodes that handle block mining, through next-generation PoW consensus
- 3.Storage Nodes that handle block writing, adding new blocks to the historical ledger
The compute nodes package transactions into blocks and send the blocks out to the miners. When the miners have completed their mining consensus along with all verification checks, the compute nodes add the mined block to the blockchain by sending it to the storage nodes. The compute nodes and storage nodes interact with each other via a RAFT consensus mechanism in order to maintain a single chain. At every stage, any decisions made by the compute nodes (e.g., selecting transactions to package) are fully and verifiably fair, employing uncontestable randomness that can be checked by any and all nodes in the mining network.
This specialisation allows for transaction block handling, validation and processing to be optimised, taking PoW from a few transactions per second to thousands. The use of sub-networks introduces a novel mechanism to enable compliance without introducing control. Data privacy, trade compliance and service levels as programmable elements introduced through a sub-network with fair, transparent governance properties have the potential to save enterprise significant time and cost while dramatically reducing risk.
Proof-of-Work is not without its problems. One of those is the knock-on effect of greed and the technological arms race in developing faster mining equipment, leading to a highly unequal and increasingly centralized network in terms of mining power. The table below demonstrates the sheer scale of the inequality in the distribution of hashrates. Single machines can reach 8 orders of magnitude greater hashrates than a standard CPU. When thousands of such machines are combined together in parallel (so-called ASIC 'farms'), the ability to exert control over the blockchain becomes a real threat to the secure operation of the ledger; moreover, the political power wielded by such large centralized bodies over the blockchain protocol and its development can be substantial.
The vast range of hashrates among mining technology
To address the problem of extreme inequality in mining, we borrow from the natural world and the laws of thermodynamics for a system of interacting particles. In the figure below, red represents a particle with a high temperature, while green represents a particle with low temperature. In such a system of particles that are in thermal contact, the rate of heat transfer proceeds as a function of the difference in temperature between each particle pair.
In nature, heat flows between particles at a rate proportional to the temperature difference
In a blockchain network, decentralization is arguably the property that gives a blockchain its power and its utility. Inequality in hashrate is a threat to decentralization, and therefore should be minimized. To this end, one can imagine a blockchain network as consisting of 'particles' (in this analogy, mining nodes) of different temperatures corresponding to the hashrates of the nodes, and apply a similar approach as shown above to bring the (effective) hashrates closer to 'thermal balance' i.e. closer to a homogeneous distribution. This can be done peer-to-peer by determining the appropriate 'heat' transfer and then employing an individual mining difficulty to modify the effective hashrate of each miner.
The full technical description of this balancing protocol is the subject of a separate paper, but briefly, the block processing power of the network (the ‘node temperature’) is moved towards a homogeneous distribution by decreasing the effective hashrate of the ‘hot’ nodes (e.g., the ASICs) and increasing it for the ‘cold’ nodes (e.g., the CPUs). This is done smoothly, allowing our node temperature quantity η to change between nodes pairwise using a simple differential equation, which for miner A takes the form:
where ∆η ≡ηA−ηB, and α is a normalisation constant. As a result, the algorithm avoids a centralised controller and operates autonomously.
The Zenotta blockchain network functions in exactly this way. A secure, peer-to-peer algorithm adjusts the iterative difficulty of the hash function in order to balance the effective hashrate between a pair of miners, and this adjustment is further checked and verified by other miners. The degree of balancing for the network is a tunable parameter, allowing the optimum to be found that ensures high efficiency, security, and in particular, inclusivity for those miners with less powerful machines that find themselves left behind in the standard PoW 'race to the top' that occurs in mining chip manufacturing.
The spectrum lying between the two extremes of the homogeneous and proportional models
In this way, decentralization, the primary goal of any public blockchain, is optimized. The network is no longer dominated by monopolistic forces that threaten the security of the consensus algorithm. While a fully balanced distribution of hashrates is very much the opposite extreme to the current high level of inequality in mining power (and would cause its own problems) a sensible middle ground somewhere on the spectrum between the standard 'proportional model' of Bitcoin and the 'homogenous model' that is the theoretical end state of the balancing algorithm would be a win-win scenario for a PoW network.
If the bitcoin price goes up by 10x, you would expect the energy consumption of the network to also go up by 10x. (Christopher Bendiksen, CoinSHARES report on BTC mining, 2019)
The problem of greed in Bitcoin mining is considerable. One of the knock-on effects of the scramble to out-perform other miners and secure a larger piece of the payout has been a huge increase in the energy usage and carbon footprint of the Bitcoin blockchain. The figure below shows the rise of the energy consumption of the Bitcoin network over a 4.5 year period.
Bitcoin historical energy consumption in annualized TWh. Figure credit: LongHash.com
We wish to stress that the public perception and understanding of this energy consumption trend and its extrapolation into the future is often based on faulty reasoning and misinformation -- for a more balanced discussion we recommend this medium post. However, no-one can predict for certain how much more the fierce competition for mining dominance will fuel the carbon footprint of Bitcoin (and other PoW blockchains) and of course, finding greener solutions to resource-intensive industries is vital in the fight against climate change.
To this end, Zenotta's next-generation Proof-of-Work consensus protocol uses the balancing of effective hashrate to create a green mining network. Although seemingly counterintuitive due to the fact that Proof-of-Work was designed to be energy-intensive, the reduction of a miner's ability to be greedy and take a far larger slice of the total network hashrate has the effect of keeping the energy consumption to a minimum (several orders of magnitude lower than traditional Proof-of-Work consensus). At the same time, miners are not disincentivised to make use of dedicated, energy-efficient chips such as ASICs, but instead are limited in the fraction of the block time that their highly effective machines can crunch the hash. This allows the mining network to grow ever more efficient, as better and more powerful mining rigs are developed, while preventing this arms race from affecting the energy consumption of the network.
The balanced nature of the mining network has implications that go beyond environmental benefits and establish a fairer and more inclusive economy for Smart Data. The on-ramp to PoW mining is incredibly steep, and all but a few with the most resources are precluded from joining. The Zenotta consensus algorithm allows the arms race that develops faster and more efficient processors to happen but prevents it from muscling out those with average CPUs. With a network where even the least powerful miners can mine profitably, the responsibility for securing the blockchain can be shared by all. Through mining and helping to secure the network, individuals gain further ability to create Smart Data and participate in the Smart Data economy. They also gain voting rights on aspects of the network protocol, which improves the democratic strength of the ecosystem.
The mining network is the backbone & the facilitator of the transactions made on a blockchain. A strong and fair data economy needs a strong backbone, and a fair means of operation, and so the use of the long established, tried & tested consensus mechanism -- Proof-of-Work -- coupled with a protocol designed to balance mining power and reduce inequality, inspired by the laws of thermodynamics, provide respectively the required strength and the required fairness.