Recently, Vitalik and MakerDao founder Rune had a dispute over the new chain plan in MakerDAO’s final plan.

Rune proposed to the MakerDAO community to use the Solana/Cosmos solution to deploy a new chain. Vitalik disagreed with this and used the action of selling 500 $MKR in the secondary market to express his attitude to the encryption market and community.

Many people in the encryption industry and community were also confused and puzzled by Rune's proposal, although he later provided some additional explanations.

In particular, the idea of ​​using Solana as an alternative to the new chain has attracted more and louder opposition. Everyone does not understand the final plan, which aims to achieve complete decentralization of the MakerDAO protocol. Why should we choose this one with the label of centralization (Solana bears the burden of The cross-branded) public chain is used as a new chain solution.

This article will strive to avoid falling into the discourse system of ideological and politically correct discussions in the crypto world, and choose to understand decentralization and centralization from the perspectives of slime network, random graph centrality and the topology of Solana main network, and to understand Rune's proposal.

Slime mold network: decentralization under resource redundancy vs. centralization under resource scarcity

Many people in the crypto world, including me before, have an obsession and conjecture about decentralization: decentralization is just and in line with nature; centralization is evil and goes against nature.

Then we will start by observing nature to understand centralization and decentralization. One of the most suitable objects for our observation is the slime mold that has both centralized and decentralized structures.

Slime molds are a group of microorganisms that are often classified as fungi. However, unlike traditional fungi, slime molds at some stages have a single-cell protoplast structure (decentralized) rather than a multicellular fungal body (centralized).

The life cycle of slime molds consists of two main stages: the vegetative stage and the reproductive stage.

Vegetative stage: In this stage, slime molds exist as single cells and obtain nutrients by absorbing organic matter (such as bacteria, algae, fungi, etc.). They exhibit a unique movement behavior in the process of searching for food, usually moving forward with cytoplasmic flow or telescopic movement.

Reproductive Stage: When slime molds' environmental conditions deteriorate or their resources are depleted, they enter the reproductive stage. At this stage, many single-celled slime molds will assemble together to form a large multinucleated cell body, often called a "fruiting body" or "accumulator". This accumulation eventually divides into multiple spores, which spread to new environments and begin a new life cycle.

To put it simply, when resources are redundant, each single cell in the slime mold network is an independent individual and cooperates to survive in a decentralized manner; while when resources are scarce, each single cell in the slime mold network Cells will specialize into specific functional cells and cooperate in a centralized manner to survive.

Both decentralization and decentralization are natural structures. They are just an adaptation of the slime mold network to the distribution of external resources. However, the centralized system prioritizes overall efficiency, while the decentralized system prioritizes individuals. Equity is a priority.

In my opinion, the various decentralized and centralized architectures of the blockchain mainnet in the crypto world are also an adaptation to the distribution of external resources, but what the slime mold network needs is water. and sugar, while the blockchain mainnet needs funds, users and developers.

In the entire crypto world, resources such as funds, users and developers are not evenly distributed, but present a typical power law distribution. The Bitcoin ecosystem and the Ethereum ecosystem almost monopolize more than 80% of the resources for the Bitcoin and Ethereum networks. In other words, the security and trustless network features and fair narrative brought about by redundant decentralization are much more important than efficiency, scalability and high TPS, so their degree of decentralization is higher than that of other L1 public chains.

As latecomers, other L1 public chains, in order to adapt to the external environment of scarce funds, users and developer resources, actively choose to pursue efficiency, scalability and high TPS in network structure design, such as Solana. In fact, this decentralization and The centralized adaptive process not only occurs between the Bitcoin and Ethereum networks and other L1 public chains, but also occurs within the Bitcoin and Ethereum networks.

At the beginning of the mainnet launch of Bitcoin and Ethereum, block space and block reward resources were extremely redundant. The mainnet was highly decentralized and blocks were evenly distributed among nodes.

But as time goes by, more and more nodes and computing power join the competition for block space and block rewards on the Bitcoin and Ethereum mainnets, so mining pools begin to appear, and the degree of centralization of Bitcoin and Ethereum increases. Higher and higher.

The Bitcoin mainnet has even seen a single mining pool own more than 31% of the computing power, and the Ethereum community is now arguing over the fact that a single entity in Lido controls more than 30% of the staking rights.

To summarize, by observing the slime mold network, we can discover the basic fact that decentralization and centralization are both adaptations of the network/system to external resource constraints, and they are both natural.

Centrality of a random graph: the probability of a node connecting to other nodes determines the degree of decentralization

Random Graph Centrality is a measurement method used to analyze the importance of nodes in a network, and is usually used to study the behavior of nodes in random graph models.

It is different from traditional network centrality measures (such as degree centrality, betweenness centrality, and closeness centrality) because it focuses more on the position and influence of nodes in the random graph model.

In a random graph model, the topology of the network is usually randomly generated, and the connections of nodes and edges are random. This model can be used to study properties in some real networks, such as social networks, biological networks, or Internet topology.

Now we use the random graph model to briefly analyze the decentralization and centralization of the encryption world.

The ideal decentralized network in the crypto industry is a uniformly distributed random network with no central node. Each node is connected to the same number of other nodes, and its degree centrality is 1.

However, the generation probability of this uniformly distributed random network in the random graph model is very, very small. In other words, in a sense, the edge generation probability setting greatly affects the degree of decentralization/centralization of the network. In Solana In the system concept, the edge generation probability corresponds to the concept of Fanout. In the newly deployed mainnet version 1.14, Solana has adjusted the Fanout mechanism in order to enhance the stability and scalability of the mainnet.

To summarize:

1. An ideal decentralized network with degree centralization of 1 has a very low probability of appearing under natural random conditions.

2. Under natural random conditions, the degree of decentralization of the network is determined by the edge generation probability. The closer the edge generation probability is to 1, the higher the average degree of decentralization of the randomly generated network.

Solana mainnet topology: layering and fan-out

Solana's self-description is: Solana is a public chain using the new development language Rust, with high scalability and performance. Its design goals are to achieve high TPS (transaction processing per second), use the Rust programming language, and low gas costs. And excellent scalability to make up for or even replace the shortcomings and status of Ethereum.

Refined into two key points:

High scalability. For example, the slime mold network is an adaptation of Solana to an environment where funds, users, and developers are scarce, and it is a survival strategy that pursues efficiency;

Ethereum challenger. A common marketing technique;

Solana's high TPS (transaction processing per second), use of the Rust programming language, and low gas fees are all designed to improve the efficiency of the system and compete with other L1s for the scarce resources left by Bitcoin and Ethereum in the encryption ecosystem.

Of course, the above characteristics of Solana arise from its network structure. Solana’s consensus mechanism uses Tower BFT, which combines the concept of PoH (Proof of Time History) clock and Gulf Stream.

Solana's propagation engine is Turbine, which consists of two parts: Erasure Batch Construction & Transmission (the construction and transmission of erasure code batch processing) and Turbine Path (Turbine path). Turbine Path can be regarded as the Solana main network. The topology diagram of version 1.14 is shown in the attached figure.

The typical characteristics of the Solana mainnet topology are layering and fan-out

In the Tx propagation network, the Solana main network will divide the nodes into several layers, with the Leader node serving as the initial sending node, and the remaining nodes will send Tx to i nodes in the next layer according to the parameter i set by the fanout.

Therefore, according to the analysis of the push-up random graph model, the degree of decentralization of the Solana main network changes dynamically, and the average degree centrality of network nodes is infinitely close to the current fan-out parameter setting 3.

Summarize

1. Decentralization and centralization are an adaptation of the system/network to external environmental resource constraints. Without the constraints of the external environment, it is meaningless to talk about decentralization and centralization;

2. Due to the current high scarcity of block rewards and block space, the internal ecology of the main networks of Bitcoin and Ethereum also presents a centralized structure;

3. The random graph model tells us that the closer the connection probability between the main network node and other nodes is to 1, the higher the degree of decentralization;

4. The degree of decentralization/centralization of Solana’s main network changes dynamically and can be adjusted, but the resource-scarce environment in which Solana operates does not support its pursuit of decentralization;

5. Rune chose Solana as an alternative for the new chain. First, Solana natively supports node staking but Ethereum L2 does not yet support it. Second, the degree of decentralization/centralization of the Solana network can be adjusted. MakerDAO can completely set a high decentralization level. Centralization degree parameters to meet the needs of decentralized governance of new chains.