Hajautetun interledger bridgen (DIB) suoritustehon optimointi

Abstract

All Distributed Ledger Technologies (DLTs) are subject to the "blockchain trilemma”, which states that the level of decentralization, security, and scalability cannot all be maximized at the same time. Thus, no single DLT can fulfill the requirements of all the use cases. However, linking multiple ledgers by an Interledger can help overcome this limitation. Unfortunately, there are three key problems affecting Interledgers: 1) they are solely focusing on transferring funds, 2) are only compatible with certain types of ledgers, and 3) are too complex (e.g. require changes to the ledgers to be connected), and most current solutions suffer from at least one of them. As a concept, bridging type Interledgers are the most promising approach, and specifically, the Decentralized Interledger Bridge (DIB) addresses all three problems. It is a lightweight solution enabling general-purpose atomic data transfer across heterogeneous distributed ledgers without having any restrictive assumptions on the client ledgers or requiring any changes to them, and the decentralized architecture provides trustworthiness and resiliency to the solution. However, the DLT used for coordinating the multiple nodes of the bridge creates a throughput and scalability bottleneck. This thesis optimizes the throughput and scalability of DIB with a new design that eliminates the coordinating DLT while still preserving the decentralized architecture for increased trustworthiness and resiliency. The results show that the throughput per node of the new design is an order of magnitude higher and the design also scales close to linearly as a function of nodes in the bridge, while the failure tests show that it maintains essentially the same resiliency and robustness as the original design.