[2017-03-11]A Better Scaling Solution Than Segwit? Sergio Says So

Bitcoin.com recently reported on RSK Labs’ scaling solution Lumino and the whitepaper for the Lumino Transaction Compression Protocol (LTCP). The plan promises to scale Bitcoin to 100 transactions per second (tps) all by itself, and be part of a much larger scaling solution. Today, RSK Labs officially announces the development and testing of LTCP. The company claims the protocol can enable billions of Bitcoin users to make payments over Lumino’s off-chain payment network.
According to today’s announcement, the experimental technology will allow for scaling RSK Lab’s Bitcoin-enhancing infrastructure to reach levels of “up to 2,000 on-chain tps” (a 667x improvement over the current max tps) “and more than 20,000 off-chain tps” when the full sidechain solution is complete.

If all goes according to plan, that would let Bitcoin process 7,330 times the number of transactions that it handles now, without raising fees. According to the RSK team, “this puts RSK on a level closer to traditional financial institutions, and it can enable fair access to financial services for the 50% of the global population currently outside the financial system”.

Bitcoin.com spoke with Sergio Demián Lerner, RSK Labs’ Chief Scientist, about Bitcoin scaling solutions to find out exactly how RSK Labs’ solutions and other projects can work together.

The Path to a Billion Users

Bitcoin.com (BC): Could you explain exactly what the differences are between Lumino and the LTCP protocol?

RSK Labs’ Scaling Solutions for Bitcoin - An Interview with Sergio Lerner
Sergio Lerner
Sergio Lerner (SL): Lumino is a hub-and-spoke network that uses off-chain payments similar to the Lightning Network (with some minor enhancements). However, Lumino relies on the LTCP protocol (Lumino Transaction Compression Protocol) as an underlying protocol layer. This protocol layer radically increases the number of settlements and payment channel top-ups that on-chain transactions can handle.

Compression is achieved by a “delta” method, with additional compression by signature aggregation. LTCP can actually create “on-chain” payment channels for frequent topping-up and settling of off-chain channels.

If an on-chain payment channel is heavily used, it’s possible that an on-chain transaction consumes as low as 5 bytes of blockchain space. The benefits are huge:

We expect Lumino running over LTCP over RSK over Bitcoin to be able to serve one billion people in the upcoming years. Almost the number of unbanked people in the world.
Complementary Scaling Proposals

BC: If Segwit is implemented in Bitcoin first, does that mean Lumino would offer an average of 210 on-chain tps?

SL: Segwit and Lumino are two complementary scaling proposals. Segwit scales using a trick to soft-fork an increase of capacity, instead of hard-forking it. It’s not a scaling solution only: Segwit brings other improvements. I will support both scaling via Segwit and via other means. All things equal, LTCP scales more because the account-based ledger model uses 3 times smaller transactions, and because LTCP removes old signature data, recovering almost 64% additional space. Therefore:

LTCP scales Bitcoin 8.3 times, while Segwit only 1.8 times.
BC: How does Lumino compare to the Lightning Network?

SL: If we compare the designs of Lumino over LTCP over Bitcoin (without RSK) and the Lightning Network over Segwit over Bitcoin, Lumino is clearly a much more powerful and elegant solution. However, a Lumino/LTCP soft-fork has not been implemented, while I understand that the Lightning network has been fully implemented, so:

I will support the Lightning Network until any better solution is at hand.
BC: Could Lightning work alongside Lumino, scaling Bitcoin the rest of the way to mainstream-level tps?

SL: Lumino IS a lightning-like network, so the question doesn’t make sense. Both Lumino and the Lightning-network allow millions of payments per second. The problem is how many active users can be served, not the actual tps, because most transactions are sent end to end, not broadcast.

According to Lerner;

“The Lumino/LTCP/RSK/Bitcoin combo can serve 1 billion users (approximately)
The Lumino/LTCP/Bitcoin combo can serve 50 million users (approximately)
The Lightning/Segwit/Bitcoin combo can serve 3 million users (approximately)
The Lightning/Bitcoin combo can serve 1 million users (approximately)”

Other Scaling Solutions

BC: Are there any other solutions out there that could fill in the gap for the other 6.3 billion?

SL: It’s only a matter of time: eventually a decentralized system would be able to serve 6.3 billion users. Meanwhile, centralized systems like Coinbase or Xapo can make payments between their own users as simple as updating their own private ledger records.

BC: Some scaling solutions sound like they would compete with each other. For instance, once RSK Lab’s sidechain solution is deployed, would Blockstream’s sidechain product be a competitor to it, or are they complementary?

SL: I think they would be complementary. At this point is too early to pick a winner: we must explore all paths to scalability and let the future users decide. We must learn from each success and failure.

Sidechains Optional

BC: Does Lumino use a sidechain and for what purpose?

SL: The LTCP paper focuses on using Lumino over LTCP over the RSK sidechain over Bitcoin, because RSK allows very fast block confirmations, and because LTCP can be easily implemented over an account-based ledger, such as RSK. Reaching a consensus over a Bitcoin soft-fork can be extremely hard, while improving Bitcoin by using a sidechain does not require one to ask permission to anyone, nor can sidechain compromise Bitcoin security in any way.

Being a sidechain, RSK is not bound to the current block size status quo. We can create larger blocks (or more frequent smaller blocks). RSK can make Bitcoin scale without compromising Bitcoin decentralization, if the RSK node resource consumption is kept low (even if it’s higher than Bitcoin). We expect a large number of Bitcoin users to participate.
However, Lumino and the LTCP can also be implemented directly as a Bitcoin soft-fork, which allows an increase of number of transactions per second in Bitcoin to 100 tps or more (the exact figure depends on the usage patterns), using extension blocks. Also, a Lumino/LTCP soft-fork would allow approximately 50 million users to be part of the Lumino/Bitcoin network, but as a comparison, the Lumino/RSK combo can reach one billion people.

BC: In a tweet a few weeks ago, you said: “I know that Bitcoin can scale to serve 400 million users and I won’t need to ask permission to anyone to scale it.” Is deploying Lumino as a soft fork still permissionless?

SL: I was referring to Lumino over RSK: when you create a sidechain, you don’t have to ask permission from anyone. People will use it or not, but you can deploy it yourself. Soft forks do require 51% of the miners to enforce the soft fork rules.

In Response to Lumino’s Critics

BC: The LTCP paper appears to say that Lumino will use the OP_RETURN field in Bitcoin transactions, and therefore not allow for multisignature transactions, nor business use of Bitcoin for blockchain timestamping, is that correct?

SL: LTCP plus an extension to the smart contract signature verification system enable multisignature transactions. I speak about this extension in the MIT presentation that is online. There is no limitation.
link:https://news.bitcoin.com/a-better-scaling-solution-than-segwit-sergio-says-so/