Overview

Currently when a node is syncing via p2p the da retrieval system is still looking through da heights sequentially. The assumption here is that p2p sync will be much faster than da sync creating a large backlog of blocks that are not confirmed.

Proposal

We should provide da ranges or specific heights from p2p sync in order to go directly to the height the data is included in instead of searching.

There are two paths:

First, the path in which another node is syncing with the head of the chain, in this case the da inclusion height is not known to the sequencer which should cause the full node to sequentially search the heights as they appear.

Second, the path in which another node is syncing after a block has been included, in this case the aggregator has stored the height the blob was included in. For this case we can include the included DA height into the gossiped data allowing the syncing full node to request blocks from celestia-node directly instead of sequentially requesting every block.