To increase the network's capacity in monolithic networks, when you increase the performance of your hardware, you can expect similar improvements on your throughput.
However, with a modular stack, given that you can have a single node for execution at the execution layer and trust a single honest node assumption for security, you can potentially use supercomputers there and abstract away the execution bottleneck.
Remember that the DA layer, which you need to post data to maintain security, can scale with the square root with the data size.
In the monolithic stack, however, you would need the whole network to run supercomputers (can't trust on a specialized high throughput DA layer), which at that level, the economics would be impossible to sustain.
Hello! Can you explain why investments in hardware result in "meagre linear growth"?
To increase the network's capacity in monolithic networks, when you increase the performance of your hardware, you can expect similar improvements on your throughput.
However, with a modular stack, given that you can have a single node for execution at the execution layer and trust a single honest node assumption for security, you can potentially use supercomputers there and abstract away the execution bottleneck.
Remember that the DA layer, which you need to post data to maintain security, can scale with the square root with the data size.
In the monolithic stack, however, you would need the whole network to run supercomputers (can't trust on a specialized high throughput DA layer), which at that level, the economics would be impossible to sustain.
- Cem Ozer