Payment Channel Networks for Blockchain Scalability
Status: concluded Period: November 2015 – May 2021 Funding: About 69,500 € Funding organization: Tim Spa (from 2019, Politecnico di Torino) Person(s) in charge: Marco Conoscenti
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
Bitcoinis 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
2021
Marco Conoscenti; Antonio Vetrò; Juan Carlos De Martin
@article{<LineBreak> 11583_2907592,
title = {CLoTH: A Lightning Network Simulator},
author = {Marco Conoscenti and Antonio Vetrò and Juan Carlos De Martin},
url = {https://doi.org/10.1016/j.softx.2021.100717},
doi = {10.1016/j.softx.2021.100717},
year = {2021},
date = {2021-06-04},
urldate = {2021-01-01},
journal = {SoftwareX},
volume = {15},
publisher = {Elsevier BV},
abstract = {Payment-channel networks are one of the most promising solution to the well-known issue of blockchain scalability. In this work we present CLoTH, a simulator of the Lightning Network — the mainstream payment-channel network, used in Bitcoin. CLoTH simulates the execution of payments in a payment-channel network and produces performance measures such as the probability of payment success and the average payment time. To the best of our knowledge, CLoTH is the only simulator that faithfully reproduces the Lightning Network code functions, and this ensures the reliability of simulation results. In this work we provide a detailed description of the new, refactored, publicly-usable version of CLoTH, and we show simulations on the multi-path-payment feature, a recent Lightning Network feature that aims to minimize payment failures.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Payment-channel networks are one of the most promising solution to the well-known issue of blockchain scalability. In this work we present CLoTH, a simulator of the Lightning Network — the mainstream payment-channel network, used in Bitcoin. CLoTH simulates the execution of payments in a payment-channel network and produces performance measures such as the probability of payment success and the average payment time. To the best of our knowledge, CLoTH is the only simulator that faithfully reproduces the Lightning Network code functions, and this ensures the reliability of simulation results. In this work we provide a detailed description of the new, refactored, publicly-usable version of CLoTH, and we show simulations on the multi-path-payment feature, a recent Lightning Network feature that aims to minimize payment failures.
@article{<LineBreak> 11583_2757472,
title = {Hubs, Rebalancing and Service Providers in the Lightning Network},
author = {Marco Conoscenti and Antonio Vetro and Juan Carlos De Martin},
url = {https://ieeexplore.ieee.org/document/8839024},
doi = {10.1109/ACCESS.2019.2941448},
year = {2019},
date = {2019-09-16},
urldate = {2019-01-01},
journal = {IEEE Access},
volume = {7},
pages = {132828–132840},
publisher = {IEEE},
abstract = {Payment channel networks are the most developed proposal to address the well-known issue of blockchain scalability. Currently, the Lightning Network (LN) is the mainstream and most used payment channel network. In a previous work we introduced CLoTH, a payment channel network simulator we developed to analyze capabilities and limitations of such networks. In this work, using CLoTH, we present results of three groups of simulations on a recent snapshot of the LN, aimed to shed a light on the following aspects. Firstly, we investigated how hubs influence the LN performance. Then, we analyzed the effectiveness of two different channel rebalancing approaches, an active and a passive one. Eventually, we studied performance of the LN when a few service-providers nodes receive payments from the other network nodes, which is a typical use case of payment channel networks. We found that the LN is resilient to the removal of hubs, that our passive rebalancing approach reduces of about one fourth the payments failures due to channel unbalancing, and that in the service-providers scenario a consistent part of payments fails because channels directing to the service providers become unbalanced. Our work contributes to prove the strengthen of the Lightning Network when removing hubs and its weakness in the service-provider scenario. In addition, the passive rebalancing approach proposed in this work is a good candidate for the implementation in the Lightning Network protocol to mitigate channel unbalancing.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Payment channel networks are the most developed proposal to address the well-known issue of blockchain scalability. Currently, the Lightning Network (LN) is the mainstream and most used payment channel network. In a previous work we introduced CLoTH, a payment channel network simulator we developed to analyze capabilities and limitations of such networks. In this work, using CLoTH, we present results of three groups of simulations on a recent snapshot of the LN, aimed to shed a light on the following aspects. Firstly, we investigated how hubs influence the LN performance. Then, we analyzed the effectiveness of two different channel rebalancing approaches, an active and a passive one. Eventually, we studied performance of the LN when a few service-providers nodes receive payments from the other network nodes, which is a typical use case of payment channel networks. We found that the LN is resilient to the removal of hubs, that our passive rebalancing approach reduces of about one fourth the payments failures due to channel unbalancing, and that in the service-providers scenario a consistent part of payments fails because channels directing to the service providers become unbalanced. Our work contributes to prove the strengthen of the Lightning Network when removing hubs and its weakness in the service-provider scenario. In addition, the passive rebalancing approach proposed in this work is a good candidate for the implementation in the Lightning Network protocol to mitigate channel unbalancing.
@article{<LineBreak> 11583_2712593,
title = {The CLoTH Simulator for HTLC Payment Networks with Introductory Lightning Network Performance Results},
author = {Marco Conoscenti and Antonio Vetrò and Juan Carlos De Martin and Federico Spini},
url = {www.mdpi.com/2078-2489/9/9/223/pdf},
doi = {10.3390/info9090223},
year = {2018},
date = {2018-09-01},
urldate = {2018-01-01},
journal = {Information},
volume = {9},
number = {223},
pages = {1–26},
publisher = {MDPI},
abstract = {The Lightning Network (LN) is one of the most promising off-chain scaling solutions for Bitcoin, as it enables off-chain payments which are not subject to the well-known blockchain scalability limit. In this work, we introduce CLoTH, a simulator for HTLC payment networks (of which LN is the best working example). It simulates input-defined payments on an input-defined HTLC network and produces performance measures in terms of payment-related statistics (such as time to complete payments and probability of payment failure). CLoTH helps to predict issues and obstacles that might emerge in the development stages of an HTLC payment network and to estimate the effects of an optimisation action before deploying it. We conducted simulations on a recent snapshot of the HTLC payment network of LN. These simulations allowed us to identify network and payments configurations for which a payment is more likely to fail than to succeed. We proposed viable solutions to avoid such configurations.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The Lightning Network (LN) is one of the most promising off-chain scaling solutions for Bitcoin, as it enables off-chain payments which are not subject to the well-known blockchain scalability limit. In this work, we introduce CLoTH, a simulator for HTLC payment networks (of which LN is the best working example). It simulates input-defined payments on an input-defined HTLC network and produces performance measures in terms of payment-related statistics (such as time to complete payments and probability of payment failure). CLoTH helps to predict issues and obstacles that might emerge in the development stages of an HTLC payment network and to estimate the effects of an optimisation action before deploying it. We conducted simulations on a recent snapshot of the HTLC payment network of LN. These simulations allowed us to identify network and payments configurations for which a payment is more likely to fail than to succeed. We proposed viable solutions to avoid such configurations.
