Skip to content
MEV Workshop at SBC 2026

Optimum Peer Turbo: A Scalable and Efficient Solution for P2P Broadcasting

Muriel Médard

Session summary

The presenter introduces Optimum Peer Turbo, a peer‑to‑peer broadcasting approach that applies random linear network coding (RLNC) to blockchain-style multicast. She frames erasures broadly as packets that arrive too late to be useful and reviews existing options: uncoded shards, structured erasure codes such as Reed–Solomon or Raptor codes, and RLNC viewed as sending random equations over original shards. She emphasizes that, for optimal multicast over arbitrary networks, RLNC is analytically shown to be uniquely optimal, while structured codes become sharply less effective as networks scale according to her cited results.

Building on this, the authors propose code‑agnostic RLNC recoding at sinks and introduce a “turbo” mechanism where receivers exchange RLNC-coded equations over weaker links. In a fluid-approximation model, this extra term is not divided by the number of receivers, so propagation speed remains high even as the network grows. The simulations described suggest that turbo recoding quickly brings nearly all nodes to decoding threshold, collapses latency gaps between accelerated and non‑accelerated nodes, and lets participants without dedicated bandwidth benefit from those who invested in faster links, with implications for fairness, misattestations, and MEV-related scalability.

Key takeaways

  • Médard argues that random linear network coding offers strong multicast properties because nodes can recode received information instead of forwarding fixed shards.
  • The proposed turbo mechanism lets receivers exchange coded information over weaker links, reducing dependence on individually provisioned fast connections.
  • Reported simulations suggest smaller latency gaps between well-connected and other nodes; practical MEV and fairness benefits remain implications of the model.