PREFACE
Hello everyone! It’s been a while since my last post. I previously promised to write an article about Blockchain, and today I finally have the chance to sit down and put my thoughts into words. My goal is to help you gain a fundamental understanding of current Blockchain platforms and their scaling solutions.
For those who are new to my blog, my articles cover Backend, Blockchain, and DevOps. I hope that by sharing knowledge with the community, I can receive constructive feedback, improve my skills, and deepen my expertise. I also write personal reflections on work and life from my own perspective.
Web 3.0 and Its Limitations
Recently, we have heard a lot about Web 3.0, which is often described as the next revolution in the evolution of the Internet. Built on Blockchain technology and distributed systems, Web 3.0 aims to give users full control over their personal data, including their identities and assets, while reducing the concentration of power in centralized data centers in a secure and transparent manner.
However, implementing Web 3.0 still presents many challenges. Issues with scalability and performance, along with a relatively complex user experience, make it difficult for ordinary users to adopt. Therefore, widespread implementation and adoption of Web 3.0 will require a long process of improvement, as current Blockchain infrastructure still struggles to meet these demands.
The Blockchain Trilemma
Over the past few years, Blockchain platforms have emerged rapidly, ranging from Layer 0 to Layer 2, each attempting to solve different problems.
However, in my view, when we look at the underlying principles, every Blockchain platform faces three competing aspects of what is known as the Blockchain Trilemma. Depending on the platform's specific objectives, trade-offs must be made between two of the following three factors:
Decentralization — Security — Scalability.
The Blockchain Trilemma
To make this easier to understand, imagine a governing organization such as a company or a country. When decision-making is highly decentralized, with many people participating in leadership discussions, decisions tend to take longer. In return, the organization may benefit from greater safety and transparency.
For example, the United States may need more time to discuss and reach a consensus on a major policy decision than China, partly because of the differences in how power is distributed and concentrated.
In this article, I will focus on scalability, which I believe is one of the biggest obstacles preventing Blockchain from becoming widely adopted and easily accessible. I will also explore how current Blockchain ecosystems are addressing this challenge.
Scalability Across Blockchain Layers
When discussing scalability, we generally approach it from two directions: vertical scaling and horizontal scaling. Blockchain platforms follow a similar principle. Scaling solutions can be implemented at different layers, depending on the ecosystem, leading to the emergence of Blockchain architectures ranging from Layer 0 to Layer 2.
So, what exactly is a layer, and what criteria can we use to determine which layer a Blockchain platform belongs to?
- Layer 0: Hardware, Network, and Data Structures
For Blockchain platforms built at Layer 0, developers focus on fundamental components such as hardware optimization, network transmission protocols, and restructuring the data architecture of a Blockchain network.
Common examples include Cardano, Cosmos, and Polkadot.
- Layer 1: Consensus
For Blockchain ecosystems operating at this layer, developers focus on improving scalability through changes to distributed consensus protocols.
Well-known examples include Bitcoin (Proof of Work), Ethereum (Proof of Stake), and Solana (Proof of History).
- Layer 2: Scaling Based on Layer 1
Recently, Layer 2 solutions have gained significant attention as an approach to increasing transaction throughput and reducing costs.
These ecosystems are built on top of Layer 1 and employ different approaches, including state channels, off-chain processing, sidechains, rollups, and optimistic rollups.
Well-known examples include Polygon, Optimism, and zkSync.
- Layer 3: DApps
There is another layer that is often overlooked but can still provide effective opportunities for scaling: decentralized applications (DApps).
At this layer, we can optimize or batch transactions and streamline the services and processing workflows involved. This helps applications scale while delivering a smoother user experience.
Solutions to the Scalability Problem
Now that we have distinguished the different Blockchain layers and their corresponding ecosystems, let's take a closer look at how current Blockchain networks approach the scalability problem.
Layer 0
A common characteristic of ecosystems at this layer is the divide-and-conquer approach. Transactions can be processed in parallel to increase throughput, measured in TPS (transactions per second).
Cardano
Cardano is one of the longest-running technology development projects in Blockchain history. It adopts a divide-and-conquer approach through its Ouroboros PoS network, which is organized into epochs and slots and incorporates sharding concepts.
However, as of the time of writing, there still do not appear to be many clear results.
Cosmos
For the Cosmos ecosystem, the core solution involves separating the DApp layer (Cosmos SDK) from the consensus mechanism (Tendermint Core BFT), while connecting the networking infrastructure through an interface called ABCI, based on the Inter-Blockchain Communication (IBC) protocol.
In simple terms, the architecture can be represented as follows:
Application (Any language) <-- ABCI (Application Blockchain Interface) --> Networking + Consensus (PoS)
In addition, the Cosmos ecosystem provides cross-chain solutions such as:
- Child chains: Zones
- Bridges: Hub — Peg Hub (Bitcoin, Ethereum)
Polkadot
Polkadot takes a different approach by building two types of chains: one for governance and coordination, and the other for running applications.
- Relay Chain: Minimizes features and focuses on governance.
- Parachains (stake DOT) and Parathreads: Support DApps.
Layer 1
Ethereum
At Layer 1, let's take a closer look at Ethereum and current on-chain scaling solutions as part of the broader evolution toward Ethereum 2.0.
Simply put, Ethereum 2.0 also follows a divide-and-conquer approach. Ethereum would transition toward Ethereum 2.0 by gradually moving its consensus mechanism from PoW to PoS while implementing sharding based on Zero-Knowledge Proofs.
Ethereum has already successfully completed The Merge, merging the Ethereum Mainnet with the Beacon Chain as a major step toward adopting PoS consensus.
Subsequent upgrades are intended to prepare the network for sharding in the future.
Other Layer 1 ecosystems worth examining include Solana and NEAR. Let's explore whether these two platforms have truly succeeded in solving the scalability problem.
Solana
Solana introduces a concept called Proof of History (PoH). In my view, PoH has a centralized bottleneck, and the system still fundamentally requires a combination with PoS.
To put it simply, Solana addresses the timestamping problem found in Ethereum:
Ethereum: Block timestamps are determined using the Unix time provided by individual miners, which is then broadcast across the network.
Solana: The leader handles timestamping through a single centralized point and automatically attaches timestamps to blocks. This reduces latency, but the network has also experienced congestion and outages.
In this case, Solana has traded some degree of decentralization for performance. As a result, it still appears unable to overcome the Blockchain Trilemma described earlier.
On the other hand, we can also consider Solana's validator count, which stood at approximately 1,800, and its TPS, which reached around 4,000.
From these figures, we can make a logical inference that there is an inverse relationship between the number of validators and TPS: increasing TPS may require reducing the number of validators.
These figures are based on data from August 2022.
NEAR
The NEAR ecosystem is built using the Rust programming language. Its scaling approach combines vertical scaling with sharding (Nightshade) to target 100,000 TPS.
However, we can also see that reducing the size of the network through vertical scaling can make it more vulnerable to 51% attacks.
In addition, NEAR provides the Rainbow Bridge and Aurora, an EVM-compatible Layer 2 solution.
Aptos / Sui
Other Layer 1 solutions developed more recently by teams with roots at Google, such as Aptos and Sui, implement scalability based on several core principles:
- Block-STM technology
- Sharding and parallel execution
- The Move programming language
- BFT consensus protocols
Layer 2
Recently, Blockchain Layer 2 solutions have grown rapidly, introducing a variety of approaches, including:
- Rollups
- State channels
- Sidechains
- Plasma
- Validium
Within the scope of this article, I won't be able to analyze each approach in greater depth. I'll save that discussion for another article in the near future.
Conclusion
In summary, I believe that no single Blockchain platform is perfect enough to simultaneously solve all three aspects of the Blockchain Trilemma. Each platform adopts a different approach, and we should evaluate these solutions based on our actual needs and the requirements of each situation.
In my opinion, revolutionizing the Internet through Web 3.0 is a promising direction. However, there is still a long way to go before the underlying infrastructure and technology can fully mature.
In the near future, I believe Layer 2 solutions are among the most viable approaches. Further into the future, I still see potential in Ethereum 2.0, given that Layer 0 and Layer 1 solutions have yet to consistently deliver the expected results in addressing scalability.
Finally, I want to emphasize that the knowledge I've summarized and accumulated over years of work is far from perfect. Errors and subjective perspectives are inevitable.
I'd welcome constructive feedback and different viewpoints in the comments so that we can learn from one another and deepen our understanding. If you found this article useful, please leave an upvote to support me in writing more articles!
Happy Coding!