Incremental decoding experiment

[coot]

Based on cardano-diffusion:demo-chain-sync, we can quantify the performance gains from incremental decoding. Currently, block-fetch codecs are not incremental (see #5114). For larger leios blocks, which have longer transfer times, incremental decoding is expected to give better performance.

[ch1bo]

We should try to run this through the demo/burst scenario and see how it impacts the Praos > Leios priorization.

[ch1bo]

We decided that a quick note in the leios-design document could be concluding this. Much like this section https://github.com/input-output-hk/ouroboros-leios/blob/main/docs/leios-design/README.md#message-latencies