Blockchain Layer2 Based Mass E-Commerce

Abstract

Blockchain technology has become a new driving force for technological innovation and has the potential to revolutionize the e-commerce industry. Hundreds of millions of users, merchants and platforms can be connected to form a blockchain-based network, enabling mass e-commerce. In mass e-commerce, transactions are settled on a distributed shared ledger. Payment issues, security issues and transaction transparency issues in traditional e-commerce will be improved. We find that blockchain Layer2 technology which moves transactions off the blockchain, has great potential to become a fast and cheap solution for mass e-commerce. We analyzed the existing Lightning Network. We found that the evolved topology leads to long path lengths and congested nodes. The routing algorithms are slow and will cause high latency for payments as well as a lot of routing-related traffic. We show that the lightning network, as it exists today, will be unable to cope with the demands of the mass e-commerce environment. This dissertation presents a blockchain Layer2 based architecture to address the payment problems of the current e-commerce industry. A tree-based topology is adapted to provide the e-commerce network among customers, merchants, and platforms. A Kademlia-based routing algorithm is incorporated to find the minimum cost path from customers to merchants. We conducted experiments on the performance of the topology and routing algorithm. Simulation results show that the network we propose can effectively shorten the path length compared with the existing unstructured Lightning Network. The K-routing algorithm we put forward can find paths faster and reduce transaction delay. The 2-Level platform node topology we design can effectively reduce the total number of payment channels required by the network. The maximum path length does not increase as the size of the network increases allowing the network to scale effectively. Furthermore, transactions can be distributed more evenly across the network of payment nodes which helps to avoid the trends of centralization exhibited by the Lightning Network. Our network also tolerates multiple node failures maintaining high availability at all times. Constructing a payment network of this scale takes some time. We show that scaling up our design to Amazon?s scale involves on-chain transactions that can be completed within days or weeks. We show that our network can be realistically deployed at a global scale. Many of the changes we propose to the Lightning Network provide a solution for future blockchain scenarios with a large number of participants. The system can scale with the demand of the network in the future. Although building the payment network involves overhead, the rewards in terms of scalability make this a worthwhile trade-of.