SCADA alarm management poses many challenges for power grid control room operators. Vast quantities of information must be managed and filtered by human operators to identify the specific alarm of interest so that system issues can be promptly addressed. The use of knowledge based systems and decision support systems to assist in industrial decision making is well established but comes with heavy cost burdens for development and maintenance. This work demonstrates a possible approach to addressing this cost issue by utilising large language models (LLMs) to parse SCADA alarms and extract information to construct a knowledge base in the form of a knowledge graph. This knowledge base can then be used to construct a decision support system suitable for control room operators to make plain language queries that assist in their problem solving. The decision support approach facilitates better and faster decision making assisting the effective management of electrical distribution networks by control room operators. The case study shown in this work demonstrates that LLMs show promising capacity for knowledge base constructions and have the potential to unlock easy access and insights from large volumes of SCADA alarm data.

The operation of power grids is increasingly complex and there is a growing need for data-driven fault diagnosis in smart grid dispatching. This study presents an innovative real-time machine learning framework that significantly enhances the detection and management of alarm events in power grid SCADA systems. At the heart of this framework is a novel deep neural network (DNN) architecture designed to efficiently identify alarm events. Additionally, an algorithm prioritizing the identification of historically relevant alarm events is developed, providing robust decision support. The proposed framework is constructed and underwent thorough testing with a comprehensive power grid dataset. The results demonstrate that the framework not only meets but exceeds the operational demands for real-time alarm detection, particularly during periods of alarm floodpeak. Moreover, the alarm event searching and decision support algorithm is proven to be a critical tool for power grid control room experts and operators, facilitating rapid decision-making by providing actionable insights based on historical and current data analysis. The successful integration of this machine learning framework into SCADA systems marks a significant step forward in the technological evolution of energy systems, leading to smarter, more efficient, and reliable power grid management.

This article proposes a blockchain-based federated learning (FL) framework with Intel Software Guard Extension (SGX)-based trusted execution environment (TEE) to securely aggregate local models in Industrial Internet-of-Things (IIoTs). In FL, local models can be tampered with by attackers. Hence, a global model generated from the tampered local models can be erroneous. Therefore, the proposed framework leverages a blockchain network for secure model aggregation. Each blockchain node hosts an SGX-enabled processor that securely performs the FL-based aggregation tasks to generate a global model. Blockchain nodes can verify the authenticity of the aggregated model, run a blockchain consensus mechanism to ensure the integrity of the model, and add it to the distributed ledger for tamper-proof storage. Each cluster can obtain the aggregated model from the blockchain and verify its integrity before using it. We conducted several experiments with different CNN models and datasets to evaluate the performance of the proposed framework.

Ensuring high data availability is a vital prerogative for Content Delivery Networks (CDN), and as we look to deploy CDN mechanisms onto mobile platforms, this imperative becomes ever more challenging. In traditional CDNs, replication ensures high availability of data, with server-loads and content-popularity often used as parameters to tightly control the process. However, the highly transient properties of such wireless and mobile devices constitute a major hurdle for any replication algorithm, rendering most simplified methods inadequate. Our contribution begins with a unique message-pulsing mechanism operating within a wireless cluster, that detects devices and ascertains their reliability. Results show the viability of our pulsing algorithm in determining a base replication level. Next, a Markovian queueing model is introduced, allowing us to induce replication based on the required speed of service. This affords finer control over the replication process, creating a more effective replication strategy suited for mobile-based CDNs. Extensive analysis of the model were performed, with parameters derived from real-world conditions. Results indicate that the model is able to compute logical values for the expected waiting times in service and thus, control the speed of replication within the CDN.

The very notion of a fully Peer-to-Peer (P2P) Virtual Environment (VE) places exacting demands on its underlying network. Subset applications such as online games are notorious for their sensitivity to latency and high bandwidth demands. By omitting the centralised mechanisms that underpin current commercial implementations, the task of managing a disparate and dynamic peer population is made ever more daunting. For any VE, arbitrating the interactions between players is an inescapable need. Online games are equal parts collaboration and competition, requiring robust conflict resolution mechanisms to govern game-play. With centralised systems, arbitration duties are simply assigned to a provisioned server infrastructure. In a P2P system however, the issue looms large. As such, managing arbitration loads across the peer population is the focus here. Being a game-play arbitrator entails added bandwidth and processing demands. Thus, great care is needed to avoid overloading peers whilst providing a responsive and uninterrupted experience. The work here exploits 3D Voronoi Diagrams (3D-VD) as a scalable, flexible and fault-tolerant P2P overlay that is able to automatically balance arbitration loads amongst peers. Simulation results indicate how 3D-VD, with the right arbitrator-selection policy, can appropriately distribute loads and reduce load fluctuations by up to 90%. This is then augmented with algorithms based on classical Newtonian gravity laws. Doing so provides an autonomous method to detect and respond to high-demand areas within the VE. Further experimentation demonstrates an ability to reduce the instances of failed arbitration attempts by 50%.

Dynamically selecting game-play arbitrators remains a key concern in fully Peer-to-Peer Virtual Environments (P2P-VE). The lack of hierarchical structures and dedicated servers mean suitable candidates must be chosen from amongst the peer population. These selected peers are elevated to the role of momentary referees, deciding on the outcomes of interactions (e.g., combat) between adversarial peers. Accordingly, a timely selection process will aid game-play responsiveness in such time-sensitive applications. There is a need to promptly affirm the arbitrator as delays will hinder peers from initiating combat, impeding game-play. The aim is to address this singular issue, investigating a myriad of selection policies contrived within the context of 3D Voronoi Diagrams (3D-VD). Prior art utilized basic 2D varieties to spatially cluster peers and limit network traffic growth. The work presented here augments this approach with a non-spatial Z-axis and the use of unique selection algorithms. Each mechanism is discussed and extensively tested, with detailed simulation results presented herein. We thus demonstrate 3D-VD's unique ability to deterministically appoint such on-the-fly adjudicators from localised candidate pools in a timely manner.

When moving towards fully Peer-to-Peer Virtual Environments (P2P-VE), the amount of network traffic generated at each peer remains a significant concern. Multiplayer Online Games (MOG) are the largest application subset of VEs and have been shown to require high frequency of update messages and minimal network latencies. Yet, this demanding criteria must be balanced with the need to also limit the otherwise quadratic growth of network traffic amongst peers. Two-dimensional Voronoi Diagrams (2D-VD) have been proposed as a way to address the inherent traffic scalability issues by naturally clustering players (and thus their update traffic) within the game-world. However, other important issues related to game-play and overall VE performance remained and were only addressed by our recent introduction of a third dimension to the VD computations (3D-VD). As our experimentation indicates, this unique approach has significant impact on the network characteristics of a P2P-VE. Due to its 3D nature, more connections are necessary per peer but a mechanism is successfully introduced to cope with the increased bandwidth requirement. More importantly, the results obtained show considerable reduction in traffic load under varying peer topologies while still maintaining the desirable features of 3D-VD.

Balancing communication workloads is a perennial performance issue in the area of Distributed Virtual Environments (DVE). The stringent time constraints of Multiplayer Online Games (MOG) complicate efforts to effectively distribute the networking load amongst servers. This issue becomes ever more exacting, when we move towards a fully Peer-to-peer virtual world (P2P-VE). We are consequently forced to factor in the limited capabilities of ordinary peers in the network. Traditional MOGs have been built on the client-server (CS) paradigm and the industry brute-force approach of over-provisioning resources is both inelegant and non-resilient in the face of failures. Moving such systems onto P2P architectures mitigates these drawbacks significantly. Our recent application of three-dimensional Voronoi Diagrams (3D-VD) onto P2P-VEs has further introduced desirable load-balancing properties in such systems. This is due to the novel use network capacity as the metric for the 3rd dimension and the subsequent use of the 3D-VD to intelligently appoint dynamic game-play arbitrators from amongst the peer population. This short paper is a preliminary report on the load-balancing properties seen in our extensive simulations. It is shown how this approach is able to appropriately distribute load in a variety of network configurations and peer populations. Thus, the performance of the collaborating peers is enhanced, ultimately leading to better game-play experience.

Voronoi Diagrams (VD) have recently been proposed as a way to enable Massively Multi-player On-line Games (MMOG) and Virtual Environments (VE) on fully Peer-to-Peer (P2P) network architectures. Using typical two-dimensional varieties, the peers maintain direct single-hop connections to others in the game-world in a scalable manner. However, direct connections do not guarantee the timely delivery of update messages amongst peers. There is also the issue of fair game-play resolution between opposing players. Thus, as is standard industry practice, game-play arbitrators are needed. Arbitrators also incorporate lag-compensation techniques which improve the responsiveness felt by individual players, thereby enhancing their experience. The approach outlined in this short paper aims to bridge the gap between industry and academia, by introducing the use of an additional 3rd dimension when deriving the VDs. Initial experimentation indicates that 3D-VD is computationally feasible within the time constraints of a typical game. A discussion into several metrics that can act as the differentiating 3rd dimension is also included.