Cosmas Ifeanyi Nwakanma

Crash-tolerant systems rely on total order protocols to ensure consistent request execution across replicated servers. The Logical Clock and Ring (LCR) protocol employs a ring-based, leaderless design that provides a high throughput but suffers latency inefficiencies under a high message concurrency due to its use of vector clocks and a fixed last-process rule for ordering concurrent messages. This paper presents the Daisy Chain Total Order Protocol (DCTOP), an enhanced version of LCR that integrates Lamport logical clocks for message sequencing and introduces dynamic last-process identification based on sender activity to accelerate message stabilisation and delivery. A modified fairness-control mechanism further balances message distribution among processes. The simulation results show that the DCTOP achieves an over 40% latency reduction compared to LCR while maintaining the same fairness and throughput across various cluster configurations.

Server replication ensures crash tolerance by enforcing a total order of input requests across multiple servers. The Logical Clock and Ring (LCR) protocol, a ring-based lead-erless total order protocol, achieves high throughput by arranging processes in a logical ring with unidirectional message flow. However, LCR design assumption may not produce optimal latency under high message concurrency due to its use of vector clocks as vector timestamps for sequencing messages and a fixed "last" process for ordering concurrent messages. To improve latency, we propose using Lamport's logical clock as a message timestamp for sequencing messages and redefining the "last" process as the nearest process in the opposite direction of message flow, ensuring a unique last process for each message sender. Fairness is preserved using a modified fairness control algo-rithm from the Fixed Sequencer and Ring (FSR) protocol. Our evaluation shows that the proposed protocol offers latency improvement better than LCR across all considered configurations. Additionally, fairness among process replicas was maintained, evi-denced by an even distribution of message sending responsibilities, with each process contributing approximately equally to total message output.

As the Metaverse progresses, its security measures must evolve to safeguard users from cyberattacks. To this end, Artificial Intelligence (AI)-powered Network Intrusion Detection Systems (NIDSs) have been implemented at the network layer to detect and respond to threats in real-time. Our survey -MetaWatch provides a comprehensive overview and analysis of relevant literature, focusing on studies that have explored NIDSs in the Metaverse. Unlike other surveys that have addressed Metaverse security issues more broadly, MetaWatch specifically provides a taxonomy of detection paradigms and defense mechanisms within Metaverse NIDS from 2021 to 2024. Additionally, MetaWatch identifies and discusses the key challenges and problems that impact the development of trustworthy, explainable, scalable, and robust Metaverse NIDS. Overall, the survey provides readers with a concise and informative summary of NIDS issues in the Metaverse and highlights open security problems worth exploring.