Developers in favor of the contentious Bitcoin Core version 30 software update cheered today when a cryptographer invented a new way to spam Bitcoin’s blockchain with images without using OP_RETURN.
After months of civil war over the proper limitations of this scripting tool, this new method showed that even reducing OP_RETURN’s datacarrier allowance to zero would not prevent all forms of arbitrary data storage.
OP_RETURN is the most popular storage area for large quantities of data unrelated to the on-chain movement of BTC. Core is fighting with a forked client of its node software, Knots, about the maximum amount of arbitrary data that the two camps will relay around the Bitcoin network’s queues of pending transactions or ‘mempools.’
Core version 30 will update the default datacarrier threshold to nearly 1MB, whereas Knots prefers a number closer to 80 bytes.
Unfortunately, as today’s discovery proves, neither setting can guarantee against all forms of on-chain spam. Specifically, the cryptographer stored a JPEG image mostly inside of a cleverly modified private key.
Spamming Bitcoin’s blockchain with images inside private keys
Indeed, BitMEX Research created a transaction containing a JPEG mostly stored within a private key—not an OP_RETURN output. Researchers deliberately used an insecure private key that could be derived solely from on-chain data on the Bitcoin blockchain itself, so that in a sense the image was stored on the blockchain.
The clever method proves that motivated users looking to store arbitrary data on the ledger can bypass OP_RETURN entirely and store data using keypairs that are consensus valid and impossible to filter out.
Read more: Bitcoin Core devs think it would be easy, and funny, to attack Knots
Although the method is unique and clever in the history of cryptography, some people hearked all the way back to Claude Shannon’s 1940s work on spam prevention, pointing out the inevitability of people evading OP_RETURN’s data filters.
The goal of Knots was never 100% spam prevention
While it is possible for a dedicated user to store spam on Bitcoin’s ledger using unconventional techniques, the Knots community has repeatedly emphasized that OP_RETURN mempool filters are an imperfect, albeit effective, deterrent against the vast majority of spam.
“100% spam prevention is *not* our stated goal – this is just a silly straw man,” repeated a leader in the Knots community. He emphasized that broad deterrence via mempool norms, not perfect prevention, is Knots’ goal of limiting OP_RETURN outputs in mempools.
“To this day, there are still people claiming that we think spam filters can stop 100% of spam,” he continued.
BitMEX Research acknowledged Bitcoin Core’s highly controversial decision to remove OP_RETURN’s data limits in version 30, scheduled for October. BitMEX Research additionally emphasized a belief that fee markets are more effective at reducing spam than mempool policy limitations.