ATQMB: Design of an augmented trust enabled QoS aware MAC model with intelligent blockchain sharding

Abstract

Enhancing security in wireless networks compromises quality of service (QoS) parameters, which reduces network scalability. In order to improve network security without compromising on QoS performance, redundant calculations must be reduced, which requires design of a performance tuning machine learning layer. Most of the research work done in this area of machine learning model design, focuses on reducing these redundancies via use of low complexity models, light weight mechanisms for trust evaluation, or simplifying key-exchange & other cryptographic processes. But these reductions in complexity also introduce security vulnerabilities into the system, for instance, using a light-weight trust model introduces trojan horse like attacks, while, using a low complexity cryptographic mechanism facilitates cryptanalysis. In order to reduce the computational complexity needed during security patch deployments, and improve overall quality of service (QoS) for wireless networks, this text proposes an augmented trust enabled QoS aware MAC model with intelligent blockchain sharding. The model works in 4 phases, wherein, initially an augmented trust model is implemented, which assists in improving QoS of network during routing. This is followed by design of a medium access control (MAC) model, which assists in further improving network QoS via an incremental learning layer that reduces packet contention, link quality issues, etc. Due to use of augmented trust layer, and incremental learning MAC model, various privacy & security loopholes are added to the network, which are handled via use of a proof-of-stake (PoS) consensus based blockchain model. The PoS model is very secure, has high traceability, but is moderately scalable, because of exponential relationship between number of blocks & computational complexity of mining. In order to reduce this complexity, and maintain a high QoS performance, this text proposes an intelligent blockchain sharding model that divides the underlying PoS blockchain into smaller shards, and manages each shard using tree-based hash range search model.