Efficient Decentralized Routing for Path-Based Transactions

Efficient Decentralized Routing for Path-Based Transactions

Settling Payments Fast and Private: Efficient Decentralized Routing for Path-Based Transactions Stefanie Roos Pedro Moreno-Sanchez Aniket Kate Ian Goldberg University of Waterloo Purdue University Purdue University University of Waterloo [email protected] [email protected] [email protected] [email protected] Abstract—1 Decentralized path-based transaction (PBT) net- the transaction partners and the transaction value [9], [13], works maintain local payment channels between participants. [14], [18], [26]. Academic and industry efforts are leading Pairs of users leverage these channels to settle payments via a towards peer-to-peer (P2P) path-based transaction (PBT) net- path of intermediaries without the need to record all transactions works such as the Lightning Network [21] for Bitcoin, the in a global blockchain. PBT networks such as Bitcoin’s Lightning Raiden Network [24] for Ethereum, SilentWhispers [11] for Network and Ethereum’s Raiden Network are the most prominent credit networks, or InterLedger [33] and Atomic-swap [10] examples of this emergent area of research. Both networks overcome scalability issues of widely used cryptocurrencies by for inter-blockchain transactions; these decentralized PBT net- replacing expensive and slow on-chain blockchain operations with works are promising for addressing scalability, efficiency, and inexpensive and fast off-chain transfers. interoperability concerns with blockchains through off-chain transactions requiring no expensive mining efforts. In fact, at At the core of a decentralized PBT network is a routing a recent blockchain event, the InterLedger team demonstrated algorithm that discovers transaction paths between sender and a transaction through seven blockchains including those in receiver. In recent years, a number of routing algorithms have been proposed, including landmark routing, utilized in the Bitcoin, Ethereum, and Ripple [36]. decentralized IOU credit network SilentWhispers, and Flare, a link state algorithm for the Lightning Network. However, the Unlike in blockchain-based PBT networks such as Ripple existing efforts lack either efficiency or privacy, as well as the or Stellar, two users u and v in a decentralized PBT network lo- comprehensive analysis that is indispensable to ensure the success cally maintain a weighted link (also called a payment channel, of PBT networks in practice. In this work, we first identify state channel, or credit link, depending on the application). The several efficiency concerns in existing routing algorithms for link’s weight characterizes the amount of funds (or assets) that decentralized PBT networks. Armed with this knowledge, we one user can transfer to the other, the exact nature of the link design and evaluate SpeedyMurmurs, a novel routing algorithm depending on the application. For instance, in a credit network, for decentralized PBT networks using efficient and flexible the weight defines the difference between the amount of credit embedding-based path discovery and on-demand efficient stabi- u v v u lization to handle the dynamics of a PBT network. Our simulation is willing to grant and the amount already owes . study, based on real-world data from the currently deployed Ripple credit network, indicates that SpeedyMurmurs reduces A PBT network builds on top of three key algorithms: the overhead of stabilization by up to two orders of magnitude routing, payment and accountability. The routing algorithm and the overhead of routing a transaction by more than a factor is in charge of finding paths with enough funds from sender of two. Furthermore, using SpeedyMurmurs maintains at least to receiver. The payment algorithm settles the funds between the same success ratio as decentralized landmark routing, while sender and receiver along the paths connecting them. Finally, providing lower delays. Finally, SpeedyMurmurs achieves key the accountability algorithm allows the resolution of disputes privacy goals for routing in decentralized PBT networks. in the presence of misbehaving users. I. INTRODUCTION arXiv:1709.05748v2 [cs.CR] 13 Dec 2017 While frequently omitted or disregarded as an orthogonal Since the advent of Bitcoin [19], many other blockchain- problem, the design of the routing algorithm is key to the based payment systems have been proposed and deployed PBT network’s effectiveness, characterized by the fraction of in practice to serve a multitude of purposes. For instance, successfully resolved transactions; efficiency, characterized by IOweYou (IOU) credit networks [3], [7] such as Ripple [1], the delays experienced during a transaction as well as the [28] or Stellar [32] leverage blockchain technology to enable overhead created by transactions; and scalability, characterized real-time gross settlements [25] between two end users across by the ability of a PBT network to maintain effectiveness and different currencies and assets significantly cheaper than the efficiency for a growing base of users and transactions. current central banking system. Ethereum [6] builds on top of a blockchain to construct a platform to run fully expressive Whereas industry supposedly considers efficiency, effec- smart contracts. tiveness, and scalability to be the main concerns for designing However, the growing base of users and transactions is a routing algorithm, we additionally emphasize the need for resulting in blockchain scalability issues [2], [21]. Moreover, privacy. Otherwise, the routing algorithm might reveal sensitive the public nature of the blockchain leads to demonstrable information such as the transaction value, the identity of sender privacy breaches of sensitive data such as the identities of and receiver, and the debt of one user to another. In this paper, we stress that all of effectiveness, efficiency, scalability, and 1This paper will appear at NDSS 2018 privacy are important to the design of a routing algorithm. A routing algorithm lacking any of these key properties is phases (approximately one per year) in the Ripple dataset unlikely to be deployed. corresponding to sudden rapid growth. • SpeedyMurmurs achieves value privacy, i.e., the value The few routing algorithms proposed so far for PBT of a transaction remains hidden, as well as sender and networks fail to achieve either privacy, efficiency, or scalability. receiver privacy, i.e., the identities of the two users remain For instance, the routing algorithm in Ripple and Stellar relies hidden from the adversary. on a public blockchain that logs the complete PBT network, thereby introducing blockchain fees and impeding privacy. In summary, SpeedyMurmurs offers an effective, efficient, Canal [35] relies on a single server to store the complete and scalable solution for privacy-preserving routing in PBT PBT network, find paths, and settle payments between users. networks, thus being a promising candidate for upcoming Therefore, the server is trivially aware of all links between deployment of such networks. Our release of the initial results users and their transactions. PrivPay [17] leverages trusted initiated a discussion about the deployment of SpeedyMurmurs hardware to encrypt the PBT network data at the server and or related algorithms in the context of the Lightning network.2 uses oblivious algorithms to hide the access patterns, thereby increasing the privacy for the links between users and their II. STATE OF THE ART AND LIMITATIONS payments. Nevertheless, PrivPay still suffers from a single point of failure and low scalability. Flare [23], a routing We first briefly overview the notion of a PBT network. algorithm for the Lightning Network, requires every user in the Then, we introduce the concepts of landmark routing and path from sender to receiver to send the current fund amounts embedding-based routing, including the description of Silent- for their payment channels to the sender, thereby leaking Whispers [11], a PBT network based on landmark routing, and sensitive information [23]. The most promising approach with VOUTE [30], an embedding-based routing algorithm, which regard to privacy is SilentWhispers [11], a decentralized PBT we adapt to PBT networks in Section IV. network without a public ledger. However, as we show in this paper, the routing algorithm in SilentWhispers lacks efficiency. A. PBT Networks In this work, we present SpeedyMurmurs, a routing algo- In a PBT network, pairs of users locally maintain links rithm for PBT networks that provides formal privacy guaran- weighted with application-dependent funds. In the Lightning tees in a fully distributed setting and outperforms the state- Network for instance, two users create a link by adding a of-the-art routing algorithms in terms of effectiveness and deposit transaction in the blockchain and update such links by efficiency. SpeedyMurmurs extends VOUTE [30], a privacy- locally adjusting their deposit’s value. The Lightning Network preserving embedding-based [20] routing algorithm for mes- thereby reduces the load on the blockchain and it has become sage delivery in route-restricted P2P networks. Targeting mes- the most promising alternative for scaling Bitcoin. sage transmission in undirected and unweighted networks The payment operation in a PBT network involves a path of rather than payments, VOUTE is unequipped for dealing intermediate users who adjust their links pairwise to effectively with weighted links and specifically changes of these weights settle funds between a sender and a receiver. In the Lightning as a result of previous transfers. SpeedyMurmurs combines Network, each intermediate user

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    15 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us