Networking, consensus, and security of blockchain applications

点击翻译按钮获取中文摘要

Blockchain technology is still far from maturity. Our proposal focuses on three crucial areas in need of further development: network performance optimization for distributed blockchain-based ledgers, consensus mechanisms, and security and privacy of data stored in the blockchain. We will investigate the impact of network characteristics on system performance, in particular on block propagation and confirmation time, probability of forking, and ledger inconsistency time. We will devise, develop, and evaluate advanced mechanisms, protocols, and message formats that will restrict the potential for inconsistencies such as forking and reduce bandwidth usage. The proposed changes will contribute to increased consistency, processing rate, and accuracy of ledger data and, thus, reduce the attack surface of blockchain-based systems. We plan to focus on hybrid consensus approaches that limit latency and guarantee finality, as is the case with centralized consensus algorithms such as PBFT (Practical Byzantine Fault Tolerance) but also offer the simplicity and ease of use in a distributed, permissionless environment, as is the case with the Nakamoto consensus used in Bitcoin. We will develop innovative mechanisms to achieve the best tradeoff between the two somewhat conflicting requirements, using the distributed paradigm with flexible membership and a rotating consensus leader. Finally, we will investigate the use of blockchain technology to develop cutting edge applications for vertical industries where security and privacy is of paramount importance, in particular in Internet of Things (IoT) and e-healthcare systems. IoT systems pose a challenge due to the large number of devices with restricted computational and communication capabilities that cannot maintain a full blown blockchain. Thus blockchain will be used for mutual authentication of IoT devices and IoT gateways or, more likely, edge or cloud servers, and also to assist in data aggregation and subsequent verification. Healthcare systems pose a big challenge on account of the large number of users and the need for individuals to control access to their healthcare data. We envisage that blockchain technology will be used to support basic security functions but not collect or store actual data due to its size. Instead, blockchain will hold user authentication information, access authorizations by data owners and (possibly) subsequent revocations of those authorizations, and (most importantly) it will record information about each and every data access that must be executed in accordance with the authorization valid at the moment of access. We will also look into sharding techniques that allow users to maintain only a portion of the global blockchain in their local storage for easy access whilst maintaining synchronism with the global blockchain.