With intense debates over the default OP_Return relay limit continuing, some have suggested OP_Return2 as a potential solution. We have summarised our understanding of OP_Return2 below.

OP_Return2 is a new proposed transaction output, that contains a hash commitment to up to 8MB of arbitrary data, but the arbitrary data itself is not required by nodes to check if a block is valid. Therefore full nodes and miners do not need to validate, propagate or store the extra data. Miners would just ignore the arbitrary data and therefore it would not contribute to mining centralisation pressure.

OP_Return2 is a softfork change to the block weight formula as follows:

8 * base data + 2 * witness data + "claimed" OP_Return2 data store < 8 million units

Special full data archive nodes can relay and store the extra arbitrary data, if they can find the data or if the data even exists. The consensus rules would only require that the amount of extra data is specified, the consensus rules don't actually require the extra data itself. A fee to the miners is then paid based on the amount of data specified in the transaction. This is a new compromise option for data.

Potential Disadvantages

• It is not clear if anyone would be willing to pay a fee for this extra data, if there is no reason for miners or full nodes to store the data. This new feature may not offer that much more than OpenTimestamps for example, but it would be much more expensive per byte.
• It might be unnecessary to do a softfork and change the weight formula, when it's already possible to put a hash of data in the Blockchain and to use a separate storage system/network.
• The extra spam data can still outbid legitimate payment related transactions and receive a steeper fee discount per byte. The size limit is now 8MB per block. Therefore, 8MB of spam images could outbid legitimate transactions, for the same cost as a 4MB inscription in the Taproot witness.