The purposes of DAL archival are:

• to be able to easily bootstrap any rollup using the DAL
• to recover some missed slots when DAL node was offline when data were shared.

The two usage are a bit different, as in one case we want a large amount of data, while in the other we need only a few slots.

• defining snapshot format

  • see what can be done with desync
  • decide what to put in the archive (slots, shards, skip-list)

• implement snapshot export

  • ensure DAL node stores network identity in it's store
  • ensure dal node stop if connected L1 node is not not the correct chain

• implement snapshot import in fresh data-dir

  • verify data
  • import data

• implement snapshot merge

  • verify data
  • merge with existing
    • incoming data erase existing data
    • existing data can be kept instead of incoming ones

• optionally reconstruct slot at export

• don't export shard if slot is already exported

• snapshot compression. Likely a ZIP file.

• implement RPCs?

  • export?
  • import?
    • with data as payload?
    • from a given DAL node (via export RPC)?