From: thepipeline_xyz
The Problem of System Instability
When a blockchain system experiences issues, common problems observed include Remote Procedure Call (RPC) failures and transactions getting reverted [00:00:23]. These issues frequently lead to explanations and excuses appearing on platforms like Twitter when something goes wrong [00:00:28]. The goal is to avoid having to make such excuses when things fail [00:00:30].
Impact on User Experience
These challenges in current blockchain infrastructure significantly impact the user experience. The speaker recounts that RPC failures and transaction reverts comprised the entirety of their user experience on Solana in the fall of 2021 [00:00:35].
Causes and Unforeseen Stresses
Such system failures often arise under conditions of extreme stress, which can be difficult to anticipate in advance [00:00:43]. Examples of high-stress scenarios include:
- NFT mints: A large number of concurrent users trying to send transactions simultaneously, such as 100,000 people attempting 100 transactions per second, can slam the system [00:00:00]. This includes scenarios where people from across the globe are “spam clicking mint buttons” [00:00:45].
- High throughput: The combination of AI agents and high throughput could yield interesting results, potentially by incentivizing AI to intentionally crash a testnet [00:00:55].
Addressing Technical Challenges and Solutions
The aim for a system like Manad is to handle these extreme stress events gracefully [00:00:07]. This involves continuously looking at data and trying to further optimize performance [00:00:11]. The ultimate goal is to see a system stand up under high stress, such as an NFT mint where everybody is slamming the system, without the need to make excuses for its failure [00:00:14]. Future testing may involve observing AI agents on the testnet to see how they interact with high throughput scenarios [00:00:50].