Chainfeeds Summary:
Last night, MegaETH completed a $20 million funding round, led by Dragonfly, with participation from Figment Capital, Robot Ventures, and Big Brain Holdings, among others. Angel investors include Vitalik, Cobie, Joseph Lubin, Sreeram Kannan, and Kartik Talwar. In a market where all participants are experiencing fatigue with public blockchain performance narratives, what will MegaETH rely on to break through?
Article source:
https://www.techflowpost.com/article/detail_18717.html
Article Author:
TechFlow TechFlow
Opinion:
TechFlow: How MegaETH Achieves Real-Time Performance: Specifically, MegaETH has only one active sequencer executing transactions at any given time. Other nodes receive state differences and update their local state via a p2p network without re-executing transactions. The sequencer is responsible for sorting and executing user transactions. However, MegaETH has only one active sequencer at any given time, thus eliminating consensus overhead during normal execution. Provers use a stateless verification scheme to verify blocks asynchronously and out of order. MegaETH has a good approach to blockchain development: measure first, then execute. That is, conduct in-depth performance measurements to identify the real problems in existing blockchain systems, and then see how to apply the node specialization approach to solve these problems within the current system. So, what problems did MegaETH identify? 1) Transaction Execution: Their experiments showed that even with a powerful server equipped with 512GB of memory, the existing Ethereum execution client Reth can only achieve about 1000 TPS (transactions per second) in a real-time synchronization setting, indicating a significant performance bottleneck in transaction execution and updates within the existing system. 2) Parallel Execution: Even with the popular concept of parallel EVM, some performance issues remain unresolved. The speedup effect of parallel EVM in actual production is limited by the parallelism of the workload. MegaETH measurements show that the median parallelism in recent Ethereum blocks is less than 2, and even merging multiple blocks only increases the median parallelism to 2.75. 3) Interpreter Overhead: Even faster EVM interpreters, such as revm, are still 1-2 orders of magnitude slower than local execution. 4) State Synchronization: Synchronizing 100,000 ERC-20 transactions per second requires 152.6 Mbps of bandwidth, and more complex transactions require even more. Updating the state root in Reth consumes 10 times more computational resources than executing transactions. In simpler terms, current blockchain resource consumption is quite high. Publicly available information suggests that MegaETH's team appears to have a Chinese background. CEO Li Yilong is a Stanford PhD in Computer Science; CTO Yang Lei holds a PhD from MIT; and CBO Kong Shuyao has an MBA from Harvard Business School and experience working at multiple industry organizations (ConsenSys, etc.). The growth manager's resume overlaps somewhat with the CBO's, and he also hails from the prestigious New York University. With four members from top US universities, the team's influence in terms of connections and resources is undeniable. Top VCs clearly favor talent from top universities.
Content source 
