Rootstock’s Chief Scientist, Sergio Demian Lerner, has been hinting about a new Bitcoin scaling solution called Lumino for over a month. Lumino’s whitepaper has finally surfaced, introducing the Lumino Transaction Compression Protocol (LTCP).
The Lumino Network and Protocol
Similar to a new envisioning of the Lightning Network, Lumino is a new type of off-chain payment channel network that plans to use compressed transactional data in Bitcoin’s blocks. Lerner describes Lumino as a new and important part of RSK‘s main platform, which is still primarily a decentralized Bitcoin sidechain, capable of an even larger amount of bitcoin transactions per second than Lumino alone.
The paper introduces a new protocol, the LTCP, which “forms the transaction layer to the Lumino network”, Lerner explains. Its purpose is to create far smaller bitcoin transactions, so as many as 100 can be processed by the network every second, which is an improvement of between 6 to 33 times today’s transaction limits.
The extension of the RSK platform was designed for “blockchains with fast block rates and account based ledgers”, including Bitcoin, but modeled after Ethereum. Lumino bridges the gap between Bitcoin and Ethereum to take advantage of the scaling benefits from the latter into Bitcoin’s blocks.
Transaction Compression Technique
Lerner describes Lumino as “a technique for transaction compression that allows processing a higher volume of transactions but storing much less information”, adding that:
Lumino transaction compression is an alternative for scaling a blockchain that achieves high compression ratios.
The lion’s share of the scaling growth comes from the compression of particular data in each transaction, while using the rest of the data as a template, which does not always need to be saved. “The critical insight behind Lumino compression is delta compression of selected fields,” Lerner wrote, “from a previous referenced transaction and the aggregate signing of previous transactions, so previous signatures can be disposed”.
If Lerner is correct, Lumino’s restructuring improvements could allow the Bitcoin network to reach 100 transactions per second (tps) “or even more, depending on the usage pattern,” he claims. “The LTCP protocol could enable the Lumino network to reach one billion active users”, he further notes. Also, once implemented, the gains should happen very quickly, without everyone upgrading and without risking a hard fork to Bitcoin’s code.
After some analysis, Lerner admits that predicting the number of active users that the Lumino network can handle is not easy, but he estimates:
Assuming each payment channel consumes a minimum of 150 bytes, and each user tops up and settles the payment channel once a month, the Lumino network can serve 1 billion active users in 4 years, as a decentralized network based on the hardware resources expected to be available for home PCs at that time (mainly higher capacity SSD drives).
This estimate is on top of any other scaling solution’s benefits such as Segwit or the Lighting Network, although it has yet to be seen if Lumino and another lightning network can work together without conflicts. After Lumino has been fully deployed, RSK’s plan for the rest of their sidechain-based product takes the network much further, in the realm of “2000 on-chain tps”, he wrote.
The largest drawback to the Lumino system may be transaction privacy. “Transaction compression means data reuse,” Lerner admits, “which means that Lumino higher compression is a tradeoff for lower privacy”. Small bits of reused data could theoretically link payments together for a user, he explains, although it is doubtful that a link between any payment and a user’s real name exist. At any rate, “the tradeoff choice is left to individual users”, he notes.
How much tradeoff depends on usage, Lerner further clarifies. “The LTCP protocol enables storage-compression but has resource bottlenecks: bandwidth usage and, to a lower extent, CPU usage”, he adds.
Another tradeoff is the use of Bitcoin’s OP_RETURN field. “To extend Bitcoins header with the fields required by Lumino”, Lerner wrote, “the fields can be stored in an OP_RETURN output of the coinbase transaction”. The controversial 80-byte field available in every Bitcoin transaction is popular with many companies and applications such as blockchain timestamping services including Factom, Blocksign, Blockstore, Stampery, and several others. If Lumino becomes part of Bitcoin, then these services may not be able to take advantage of the upgrades without re-writing their platforms to use some other method.
As a Bitcoin Soft-Fork
Like many projects nowadays, Lerner came to the conclusion that Lumino should be introduced to Bitcoin as a Soft fork, which is backwards-compatible and does not require everyone to upgrade to it in order for it to work. His paper describes what it would take to implement Lumino into Bitcoin this way. “We need to create an account address space for a sidechain embedded in the Bitcoin block, as an extension block”, he conveys. The Bitcoin nodes that do so will be able to process Lumino transactions immediately, which he believes would drastically cut down on the amount of space each transaction in a block takes up, making them as small as 8 bytes.
Interestingly, Lumino planned as a soft fork upgrade to Bitcoin “does not benefit from Segwit space savings, since Lumino exodus address is an anyone-can-spend output, the transaction that consumes this output specifies an empty script”, Lerner notes.
Images courtesy of Shutterstock and RSK