Skip to content
LogoLogo

Merged mining

A new proof-of-work chain's biggest risk is rented hashpower: Ethereum Classic and Bitcoin Gold lost millions to 51% attacks bought on hashrate markets. Arrekta borrows Bitcoin's hashrate instead: every block proves its work with an AuxPoW, a Bitcoin parent header whose coinbase commits to the Arrekta block (§9.6).

How it works

Bitcoin coinbase scriptSig:  … fa be 6d 6d ‖ reverse(aux merkle root) ‖ size u32 ‖ nonce u32 …
                                            │
                                            └─ root of a small tree of aux chains; Arrekta's block id
                                               sits at the slot chosen by (nonce, chain_id 0x0D07)
  1. The pool asks the node for work (createauxblock) and gets the Arrekta block hash and target.
  2. It puts the aux merkle root into its Bitcoin coinbase, next to any other aux chains (Namecoin, Dogecoin, Syscoin).
  3. When a Bitcoin share meets Arrekta's target, the pool submits the parent header, coinbase and merkle branches (submitauxblock).
  4. Nodes check that the coinbase is in the parent header's merkle tree, that the Arrekta block id is committed at the right slot, and that the parent header hashes below Arrekta's target.

Design choices

  • Namecoin's exact format, so pools already merge-mining Namecoin, Dogecoin or Syscoin add Arrekta with a configuration change.
  • Stricter than Namecoin: the fabe6d6d magic is required, exactly once and immediately before the root; 64-byte coinbases are rejected (merkle ambiguity); the coinbase is capped at 32 KiB.
  • One proof-of-work rule. Solo miners use a standalone parent built from the same structure. There is no second, cheaper hash an attacker could mine on.
  • Chain id 0x0D07 on every network; its low bits differ from Namecoin (1), Dogecoin (98) and Syscoin (16), so slots never collide.
  • Difficulty follows ASERT (aserti3-2d) with a 120 s target spacing and a one-hour half-life; regtest has no retargeting.

Known weakness, and the answer

Research ("Merged Mining: Curse or Cure?", ePrint 2017/791) found merge-mined chains often depend on a few dominant pools. That is why Arrekta adds a finality layer: signed checkpoints.