Payment Channel Networks for Blockchain Scalability

Status: 
ongoing
Period: 
November 2015 - November 2020
Funding: 
About 69,500 €
Funding organization: 

Politecnico di Torino

Person(s) in charge: 
Executive summary: 

The goal of the research project is to analyze capabilities and limitations of payment channel networks, the most explored solutions which aim to solve the well-known issue of scalability of the blockchain.

Since June 2017, the research project is conducted by the Nexa Center in cooperation with Fulgur Lab.

Background: 

Bitcoin is a decentralized cryptocurrency that allows mistrusting peers to send and receive monetary value without the need for intermediaries. Bitcoin relies on the blockchain, a distributed peer-to-peer public ledger which stores all the history of Bitcoin economic transactions. The blockchain protocol, which is designed to contain centralization and to allow any peer to actively participate in the system, limits the transaction throughput. For this reason, Bitcoin and other blockchain-based cryptocurrencies do not scale.

Payment channel networks are the most promising solution to address the issue of blockchain scalability, as they enable off-blockchain payments which are not subject to throughput limit of blockchain transactions. Payment channels are two-party bidirectional channels which allow the two channel parties to exchange an unbounded number of off-chain payments. Payment channels are linked together to generate a payment channel network, which allows also parties not directly connected by a payment channel to exchange off-chain payments.

The Lightning Network (LN) is the mainstream payment channel network, built for Bitcoin. At its current state of development, the LN presents critical features that, if not properly understood, implemented and controlled, might undermine the development of a healthy payment network. Some examples of these critical features are channel economic capacity, which limits payment amounts, channel unbalancing, which makes payment channels unusable in one direction, uncooperative behaviour of nodes, which may cause payment failures and lock of funds.

Objectives: 

The goal of this research project is to analyze capabilities and limitations of payment channel networks. To accomplish with this goal, CLoTH was developed, a simulator for payment channel networks. CLoTH simulates the execution of payments on a payment channel network and produces performance measures such as probability of payment failures and average payment time. CLoTH is a valuable tool to identify issues, analyze solutions and steer future developments of payment channel networks.

So far, three groups of simulations were conducted using CLoTH: (i) simulations on the Lightning Network, aiming to discover configurations of the Lightning Network in which a payment is more likely to fail than to succeed; (ii) simulations on synthetic networks generated by the simulator, to analyze the impact of the simulator input parameters on payment network performance; (iii) simulations which studied network and protocol modifications on the Lightning Network (such as removal of the hubs from the network, and protocol optimizations which address the problem of channel unbalancing).

Results: 

The main achievements of this research project are:
• CLoTH, a simulator for payment channel networks written in C. As input, CLoTH takes parameters defining a payment channel network (e.g., number of channels per node, average channel capacity) and parameters defining payments (e.g., payment amounts and payment rate). It simulates the input-defined payments on the input-defined payment network. It produces performance measures in terms of payment-related statistics, such as probability of payment failure and time to complete payments. CLoTH can be used to identify issues of payment channel networks and to estimate the effects of protocol optimizations on the network before implementing them. It can also simulate attack scenarios and can study the evolution of payment channel networks.
• Simulation results on payment channel networks. The main findings prove that the current most relevant issues of the Lightning Network are the limited channel capacities and channel unbalancing, which both cause payment failures. On the other side, simulations prove that the Lightning Network is resilient to the removal of the most connected network hubs and can support a contained level of faulty nodes. Another remarkable finding is that a rebalancing approach designed in this research work proved to be effective against channel unbalancing.
• A Ph.D, thesis entitled Capabilities and Limitations of Payment Channel Networks for Blockchain Scalability. The thesis presents the CLoTH simulator and the results of the simulations conducted so far.
• A journal paper titled Hubs, Rebalancing and Service Providers in the Lightning Network, which discusses simulations conducted on hubs, service providers and rebalancing approaches in the Lightning Network.
• A journal paper titled The CLoTH Simulator for HTLC Payment Networks with Introductory Lightning Network Performance Results, presenting the CLoTH simulator and the first simulation results.
• A literature review on the blockchain, which highlights the main research challenges of the blockchain (scalability and anonymity) and discusses possible applications of this technology. The literature review has been published as a paper titled Blockchain for the Internet of Things: a Systematic Literature Review.

Related Publications: