Although Bitcoin broke 90,000 today, there is still no obvious sentiment reversal. I joked with friends in the group that as long as BTC doesn't surpass 100,000, it can't be considered a bull run, at most just a slight rally. At least for now, the market remains uncertain, and with the upcoming holiday, it's a good time to learn about blockchain technology. Recently, by chance, I joined Sei Advocates and decided to practice with Sei @SeiNetwork.

Today, I learned about TPS ( Transactions Per Second ), which is the core indicator of blockchain network performance. It measures the number of transactions a blockchain can successfully process and confirm in one second.

To better understand the significance of TPS, let's compare some different chains:
Bitcoin: approximately 7 TPS
Ethereum: approximately 15-30 TPS
Solana: approximately 3,000-5,000 TPS
Sei Network: approximately 20,000+ TPS, even higher after the Giga upgrade

Here's a vivid analogy: this is like the number of toll booths on a highway. The more toll booths, the more cars can pass simultaneously. Now you understand why Bitcoin and Ethereum networks get congested at critical moments. When congestion occurs, the higher the gas fee you pay, the sooner your transaction gets processed.

What surprised me is that the global traditional payment network Visa has about 24,000 TPS, with peaks reaching 65,000 TPS. I always thought the blockchain industry was developing fast enough, and many high-performance public chains should have already surpassed traditional financial systems. Although Visa is centralized, it is indeed impressive. Visa can handle trillions of transactions daily worldwide without congestion—that's the kind of performance traditional finance aspires to.

Is it difficult for blockchain to improve TPS?
Theoretically, no; practically, yes. Improving TPS requires optimization across multiple dimensions such as consensus mechanisms, execution efficiency, and data structures, while balancing decentralization and security. In real operation, factors like decentralization, security, and load influence performance, which is why many public chains keep TPS conservative.

Sei Network has made many technical optimizations to improve TPS:
✅ Twin-Turbo consensus (consensus optimization)
✅ Parallel execution engine (parallel processing)
✅ SeiDB dedicated storage (data structure optimization)
✅ Native order matching (application layer optimization)
🔜 Autobahn protocol (future: multi-validator parallel block production)

This is why Sei is suitable for gaming (requiring quick response), DeFi, especially derivatives (requiring low cost and high frequency), and everyday payments (requiring cheap and stable solutions).

Of course, higher TPS isn't always better. Pursuing high performance at all costs can increase unnecessary expenses and reduce security (by minimizing validation time). In recent years, blockchain performance has indeed advanced rapidly, but to fully surpass traditional financial systems, both technology and ecosystem maturity are needed. My previous suggestion for Sei to officially develop PerpDEX and prediction markets was based on this consideration. I really hope the team can evaluate this, haha.

The attached diagram shows the TPS requirements for different application scenarios: