Original author: Justin Bons, founder of Cyber ​​Capital

Original compilation: Luffy, Foresight News

There are three ways to achieve scalability in blockchain:

• L2 expansion (BTC, ETH, TIA)

• Parallelization (SOL, APT, SUI)

• Sharding (TON, NEAR, EGLD)

Sharding is the future, and parallelization is an inevitable trend. In my opinion, this is where it all ends.

Because L2 expansion brings poor user experience and a weaker token economic model, it cannot control fees within the L1 range, and fragmentation destroys the user experience. Monolithic blockchain scaling approaches (parallelization and sharding) do not suffer from these fatal flaws because they are a coherent whole.

Parallelization is inevitable because client software would be foolish not to support multi-threading. All modern CPUs have multiple cores, but chains like Ethereum and Bitcoin still only process transactions sequentially, which results in most validator hardware being underutilized, which is a huge waste.

The same goes for shard chains, as each shard should be parallelized

• Maximize individual shard capacity

• Sharding is a further implementation of parallelization logic

• Distribute workloads across multiple machines by scaling concurrency from multiple cores

This breaks previous scalability limitations.

The sharding system is now able to achieve more than 100,000 TPS, and the theoretical limit is close to 1 million TPS. At the same time, sharding has relatively low node requirements. This is how sharding solves the blockchain trilemma.

Traditional blockchain designs face three dilemmas. Because at some point, node requirements become so high that they threaten decentralization. Since all nodes must validate all global state updates, it fundamentally does not scale. Sharding solves this problem.

Unlike traditional designs, sharded chains can scale capacity based on usage, whereas non-sharded chains will always eventually hit a cap. When a shard chain gains more usage and adoption by validators, it can launch a new shard. In other words, sharding scales linearly.

Other blockchains scale quadratically, meaning that as the network grows, the node requirements become higher and higher until physical limits are reached. There is an upper limit to what we can handle within a single silicon chip compared to what can be achieved with computer networks.

There are many wrong views about sharding. Let me make two points here:

• “You can attack a single shard”; rebuttal: since validators are random, shards share L1 security

• “No composability”; rebuttal: cross-shard communication is natively built-in, ensuring seamless interoperability

The irony of these criticisms is that L2 scaling is more likely to make the same mistakes:

• “You can attack a single L2”; this is true, especially considering that managing keys and decentralized orderers require their own consensus

• No composability; same fact, no enshrining

Fortunately, the leap from parallelization to sharding is much shorter compared to modular blockchains.

At the same time, parallelization will likely provide enough capacity for many years to come, which is why I would support the latter two options.

Monolithic expansion still allows modular expansion using L2, allowing the free market to choose the best solution; while modular expansion is more similar to the planned economy of L1 forcing modular expansion.

We should let the market choose another L1/L2.

We have to draw the line where modular blockchains are concerned, I am convinced that modular scaling is a technical dead end. Worse, it sets us back because people mistakenly associate modular design with cryptocurrency. Slow, expensive and difficult, thats what modularity is.

And monolithic designs are fast, cheap, easy to use and understand. If the community provides enough resistance, Ethereum could still return to sharding, which could lead to a block size debate-style fork as conservatives try to hang on to power.

There is no doubt that the entrenched power within Ethereum will not be easily overturned. Venture capital and tokens provide strong incentives for Ethereum’s L1 expansion. Since Ethereum also lacks good on-chain governance, it may be easier to vote with your feet.

I am not an enemy of Ethereum, but a friend of it. If Im right, then Ethereums biggest enemy is within its entrenched leadership, and the same goes for Bitcoin.

Power corrupts, and absolute power corrupts absolutely.

Tribalism aside, the bottom line is whether the evolution of blockchain technology is along the right path: as I said, monolithic scaling. Advocates of modular scaling often cite the blockchain trilemma as an argument underpinning the approach.

I respect this ideology because there are many good and smart people supporting L2 scaling.

However, this belief is based on a flawed assumption. The evidence for viable L1 expansion continues to pile up, and its becoming a mountain. It’s too big to ignore, with competing blockchains surpassing Bitcoin and Ethereum on multiple metrics.

The truth is out there, monolithic scaling is the future, enabling everyone to use the blockchain directly, bringing us back to Satoshi Nakamoto’s vision.