It’s great to see more research on Based Rollup!
For the L3 and L4 scenarios you mentioned.
- Is the use of Based Rollup more difficult compared to L2, e.g. extra latency and excessive extra workload?
- L3 and L4 will probably mostly use Validium’s DA model, which is probably a Weakly-trusted “Rollup” in general, is it necessary to use Based Rollup?