Author: @Daniel Sanchez Quiros @Thomas Lavaur
Reviewers: 🟢@Álvaro Castro-Castilla 🟢@Daniel Kashepava 🟢@Gusto Bacvinka 🟢@Marcin Pawlowski
NomosDA dispersal is the sub-protocol, used by encoders, for publishing data onto the NomosDA network. This document describes the protocol structure abstracted from the network construction that can be found in the NomosDA network specification. This document also describes the interaction between encoders and validators.
NomosDA dispersal aims to provide encoders with a simple method of distributing data into the NomosDA subnetworks. The protocol needs to be:
num_subnets
subnetworks. The network burden on DA nodes should be minimal.NomosDA dispersal is a sub-protocol for encoders delivering encoded shares of data to the NomosDA network. Encoders listed in the Service Declaration Protocol (SDP) as stipulated in the NomosDA network specification are entitled to publish data.
Encoders connected to each NomosDA subnetwork encode data blobs as per the NomosDA Encoding Protocol . Then, shares of the encoded data are distributed into each NomosDA subnetwork (0
to num_subnets-1
). Encoders wait for an acknowledgment that the encoded data was received by a DA node and publish a set of metadata to the mempool.
Optionally, encoders can perform sampling to double-check that the data was distributed properly without trust issues.
Before the dispersal protocol can be executed, the minimum number of nodes (unique ProviderId
s from declarations) must be retrieved from the SDP (Minimum Network Size). After this requirement is met, the protocol operates as described below.
The dispersal protocol starts with data to be encoded, then it works as follows: