The rapid growth of video streaming services and the increasing coexistence of multiple content providers have placed significant demands on network infrastructure. Efficient content caching is crucial to improving user experience and optimizing network resource utilization. Traditional content delivery networks (CDNs) struggle to meet the dynamic requirements of multi-provider environments. In this paper, we address the problem of multi-provider content caching by adopting fog computing to leverage localized storage and processing capabilities. To this extent, we develop CAMP (Cluster-Aware Multi-Provider Caching), a novel cooperative caching framework designed for multi-tier fog architectures. CAMP consists of two phases: resource allocation and content placement. In particular, the first phase clusters users based on location and provider affiliation, hierarchically assigns fog nodes using a fuzzy membership model, and partitions fog storage proportionally among clusters. The second phase employs a cooperative content placement algorithm that dynamically adjusts content redundancy based on popularity to ensure high availability for popular content while minimizing redundancy for infrequently accessed data. Through extensive simulation using real-world datasets, CAMP demonstrates up to 71% reduction in latency and substantial improvement in fog hit ratios, outperforming state-of-the-art caching strategies.

Machine learning (ML) has demonstrated significant promise in accelerating scientific discovery; however, its application in experimental domains is often constrained by limited data availability. In this paper, we present an Agent as a Service (AaaS) that optimizes the synthesis of electrocatalysts for hydrogen evolution, a critical process for efficient hydrogen production. The AaaS framework orchestrates multiple agents, each responsible for a distinct decision stage, to enable prediction of hydrogen evolution performance from synthesis-related features, particularly the overpotential required at fixed current densities. Random Forest (RF) and XGBoost (XGB) models are trained using five-fold cross-validation and hyperparameter tuning, enabling a comparative evaluation of predictive performance and generalization. The results highlight that careful framework design, rather than model complexity alone, is critical for achieving reliable predictions in small-data settings $(n=40)$. In particular, the AaaS framework demonstrated the efficacy of ensemble models in small-data settings, with both RF and XGB achieving high training accuracy $\left(R^{2}>0.98\right)$ and predicting consistent Ni-rich optimal regions (e.g., $N i \approx 0.64, C o \approx 0.36$). The proposed framework has the potential to be extended to a wide range of scientific and engineering problems where data is scarce but structured.

Deploying intelligent monitoring services in industrial settings often faces three simultaneous constraints. These are the absence of labeled training data, limited computational budgets that preclude GPU-based deep learning, and the need for pipelines that transfer across sensor modalities. This work reports an empirical finding relevant to all three constraints. Contrary to the expectation that photoluminescence degradation features require deep learning or expert-defined descriptors, a single unsupervised principal component analysis (PCA) component recovers approximately 44% of predictive importance in a Random Forest regressor trained on photovoltaic degradation data. We frame this as a finding about what unsupervised dimensionality reduction can recover from sensor images, together with its practical implication for monitoring service design, rather than as a new algorithmic component. The pipeline combines unsupervised feature extraction with supervised prediction, and executes on commodity hardware without GPU acceleration or deep learning frameworks. Applied to photoluminescence sensor images from only 158 samples, it achieves R2 = 0.65 and MAE = 0.29 mA under a random split. We further evaluate the pipeline under a temporal split and block-wise temporal cross-validation, and against voltage-only and intensity-only baselines, to assess how much the discovered image feature adds beyond simple metadata. The results indicate that unsupervised dimensionality reduction can serve as an effective feature discovery method for degradation-sensitive patterns in monitoring services where labeled data and computational resources are limited.

Blockchain technology is often discussed as if it emerged from nowhere, yet its architectural DNA traces directly to the decentralized computing principles James N. Gray articulated in 1986. This paper maps the conceptual lineage from Gray's requestor/server model to modern blockchain architectures, showing how his emphasis on modularity, autonomy, data integrity, and standardized communication anticipated the design of systems like Bitcoin and Ethereum, and, more recently, the Web3 movement and Layer-2 scaling architectures. We examine consensus mechanisms, cryptographic foundations, rollup-based Layer-2 protocols, and cross-chain interoperability through this historical lens, identify persistent challenges in scalability and modularity, and outline future directions toward Web4: an intelligent, decentralized internet integrating blockchain, artificial intelligence, and the Internet of Things.

Blockchain is a foundational technology for decentralised systems, but major Layer 1 (L1) networks such as Bitcoin and Ethereum face scalability and cost limits. This has motivated the development of Layer 2 (L2) solutions that process transactions off-chain while preserving L1 security guarantees. We present Agnos-L2, a flexible L2 blockchains that works across different L1 blockchains, including Bitcoin and EVM-based networks. The design separates execution, state commitment, and settlement. This allows an adaptive transaction pipeline that learns each chain’s block time, fee market, and proof rules. It then picks the best routing, batching, and proof methods. We model the system with queueing theory and optimization to study throughput, latency, and cost under security limits. We also give pseudocode for an adaptive router and a settlement scheduler. Simulations show that Agnos-L2 can reach up to 3× more throughput and cut costs by 35% in realistic network settings. It keeps L1 verifiability using succinct proofs and fraud or validity games. We compare Agnos-L2 with rollups, payment channels, and sidechains, and study its security against censorship, reorgs, and bridge risks. Our results show that one L2 design can link many blockchains while staying secure and efficient.

The proliferation of video streaming and mobile applications has intensified the demand for low-latency, high-throughput content delivery. Fog computing offers a promising solution by enabling content caching closer to end users at intermediary nodes, known as fog nodes, i.e., fog caching. However, existing fog caching strategies often fail to adapt to the rapid dynamics of user mobility and evolving content preferences, particularly on platforms like Instagram, YouTube, and TikTok. This paper presents a novel personalized fog caching strategy that proactively caches content by jointly predicting user mobility and application usage patterns. Leveraging a deep learning architecture built on transformer encoders, our strategy is to anticipate both the fog nodes users will connect to and the applications they are likely to access. It combines high-resolution spatio-temporal trajectory data with historical app usage behavior to drive fine-grained content placement across multi-tier fog networks. Simulation results demonstrate that our approach significantly improves cache hit rates and reduces retrieval latency compared to traditional popularity-based techniques.

Modern vehicles equip dashcams that primarily collect visual evidence for traffic accidents. However, most of the video data collected by dashcams that is not related to traffic accidents is discarded without any use. In this paper, we present a use case for dashcam videos that aims to improve driving safety. By analyzing the real-time videos captured by dashcams, we can detect driving hazards and driver distractedness to alert the driver immediately. To that end, we design and implement a Distributed Edge-based dashcam Video Analytics system (DEVA), that analyzes dashcam videos using personal edge (mobile) devices in a vehicle. DEVA consolidates available in-vehicle edge devices to maintain the resource pool, distributes video frames for analysis to devices considering resource availability in each device, and dynamically adjusts frame rates of dashcams to control the overall workloads. The entire video analytics task is divided into multiple independent phases and executed in a pipelined manner to improve the overall frame processing throughput. We implement DEVA in an Android app and also develop a dashcam emulation app to be used in vehicles that are not equipped with dashcams. Experimental results using the apps and commercial smartphones show that DEVA can process real-time videos from two dashcams with frame rates of around 2230 FPS per camera within 200 ms of latency, using three high-end devices.

These days, people are always connected via internet and they use map services like Google Maps all the time. In this study, we demonstrate how to take advantage of such connectivity and technologies for raising the awareness of sustainability. To this end, we develop a mobile app (The Carbon Conscious Traveller or TCCT) that helps people travel in a more carbon-conscious manner. In this demo, we showcase TCCT, which seamlessly combines navigation using Google Maps API, visualisation, and carbon footprint calculation and tracking for an extensive range of transport modes including cars, trains, airplanes and even ferries. This tool encourages people to choose environmentally friendly travel practices by giving them a concrete and accessible indicator to comprehend the impact of their activities on the environment.

To address the critical need for enhancing Quality of Service (QoS) monitoring in the logistics service delivery domain, this paper introduces a blockchain-based QoS monitoring framework that aims to automate service delivery and dispute resolution processes. Traditional QoS monitoring solutions rely heavily on human judgment and intervention, which may increase operational costs and reduce service reliability. Such behaviour highlights the demand for more efficient and transparent logistics services, which underlines the importance of utilizing blockchain technology’s immutable and decentralized nature. The proposed framework in this paper employs a graph-based approach to transform QoS requirements into Deterministic Finite Automata (DFA) format. This strategy simplifies delivery monitoring and the identification of service violations through efficient DFA traversal. By using the Ethereum network as the deployment environment, we demonstrate Traversing a DFA is computationally efficient and reduces operational costs. Extensive experiments were conducted to evaluate the cost-effectiveness of the framework, showing that monitoring a delivery using this framework costs approximately $2.59. This finding underscores the framework’s advantages in operational cost optimization compared to traditional human-based methods. Moreover, the decentralized nature of our proposed framework allows customers and businesses to jointly define and monitor QoS parameters. Therefore, a transparent and trust-based relationship between the business partners can be established.

Efficient content caching in video streaming services is important for improving user experience as well as reducing network bandwidth consumption. Fog caching combined with scalable video coding (SVC) has the potential to significantly improve caching efficiency. However, challenges such as the limited storage capacity of fog nodes and determining the optimal number of SVC layers must be overcome for their effective adoption. This becomes more complicated with multiple content providers requiring shared cache resources. To the best of our knowledge, no existing research has simultaneously tackled all these aspects. In this paper, we present Cluster-based Cooperative Fog Caching (CCo-Fog), a holistic caching strategy that enables multiple content providers to share the scarce storage of fog nodes in a multi-tier fog network to judiciously cache SVC videos in a cooperative manner. In particular, CCo-Fog consists of a cluster-based storage partitioning method and tier-wise cooperative content placement policies. The partitioning method distributes the storage of each fog node to multiple content providers for users clustered based on their population density and their proximity to fog nodes. The content placement policies determine the optimal number of SVC layers of each video for different tiers of the fog network by solving a latency-aware content placement optimization problem. Our evaluations on a real-world dataset and various configurations demonstrate the efficacy of CCo-Fog, showing a reduction in latency by about 60% and an increase in fog hit ratio by about 20% on average, compared to state-of-the-art caching strategies.

In the United Arab Emirates, housing establishments offer citizens opportunities to apply for housing loans or grants. The eligibility for a loan or grant depends on the citizen’s income. Those with incomes below a predetermined threshold can apply for grants, while those with higher incomes are eligible for loans. Along with proof of income, other information such as the citizen’s health, fitness level, number of dependents, and marital status must be submitted with the application. This information, typically sourced from various government entities, must be verified. Traditionally, verifying this information through direct contact with these entities has been inefficient and time-consuming. To address this, we propose a blockchain-based, graph-traversing framework. This framework aims to create a secure mechanism for information sharing between housing establishments and government entities, thereby automating the processing of housing support applications. The framework also represents housing regulations as a graph, enabling updates to the regulations without disrupting workflow. We conducted several experiments to analyse the cost-effectiveness of the proposed framework, finding that processing a single application costs less than $1 on average.

Efficient content caching is crucial for video streaming services, enhancing user experience and conserving network bandwidth. While traditional Content Delivery Networks (CDNs) address this need to a certain extent, fog computing is emerging as a complementary solution. In particular, this new computing paradigm utilizes nodes positioned between users and the cloud continuum (fog nodes). However, the limited capacity of fog nodes poses a challenge. Previous studies have addressed this challenge by focusing on cooperative caching, considering factors like popularity and user location, yet often overlooked the shared use of fog node storage by multiple content providers (MCPs). This paper introduces CCoFog Caching (CCo-Fog), a cluster-based cooperative content caching strategy that allocates fog node storage capacity among content providers using user clustering and the multi-tier feature of fog networks. It also proposes a content placement algorithm that considers popularity to determine the number of content copies in the network. Evaluation with real-world data shows that CCo-Fog significantly improves latency by 54% and hit ratio by 1 |6% compared to existing strategies.

We present P4Xtnd, which implements P4 programmability on resource-constrained devices with extended network functionalities for malicious device identification in IoT networks. Compared to existing methods, P4Xtnd operates in realtime and distributed manner for traffic collection, running ML at P4 Data plane with Federated Learning. Moreover, algorithms are proposed for trust assessment of sensor devices and dynamic network slicing for effective network management. With three LAN networks P4Xtnd shows its ability to detect malicious activities at higher accuracy.

Recent cyberattacks have increasingly targeted distributed networking environments like IoT networks. To detect these attacks, hidden under network traffic encryption, many centralized Machine Learning (ML) based solutions have been introduced, which are not well suited for IoT networks. This work proposes PIFL a practical approach to secure IoT networks by combining federated learning, in-network ML using P4-enabled devices, software-defined networks, and binarized neural networks. PIFL detects compromised edge devices and isolates them into separate network slices based on trust parameters derived from their behavior. We demonstrate the feasibility of PIFL using an experimental testbed with three intelligent network devices and seven IoT devices implemented on Raspberry Pi devices.

Fog computing plays a crucial role in bridging resource-constrained internet of things (IoT) devices and distant clouds by processing requests from IoT devices in a timely manner. In this paper, we present the Many-to-One (M2One) offloading algorithm taking into account mobile fog nodes. In particular, M2One identifies a coordinator fog node to distribute those requests to other fog nodes as well as cloud servers. The coordinator node is selected based primarily on computing capacity and signal strength. The actual offloading decisions are made considering computing capacity, queuing delay, transmission and propagation delay, channel status, and battery level of mobile fog nodes. Our evaluation study conducted in a simulated healthcare scenario has demonstrated the effectiveness of M2One. Results show that the proposed algorithm reduces request service delay by 35% on average, compared to that with clouds.

Blockchain, a type of distributed ledger technology, has revolutionized the digital economy such as cryptocurrencies and supply chain management with its transparency, immutability, and decentralization properties. In addition, smart contracts are introduced to the blockchain to provide programmability removing third parties for administration. Although promising, blockchains and smart contracts are closed technologies meaning they have no interaction with the external world where real-world data and events exist, i.e., off-chain data. It becomes more challenging when the off-chain data is unstorable onto the blockchain due to data volume and privately maintained by third parties for security and confidentiality. In this paper, we address the problem of enabling a private blockchain platform to access privately owned sensitive off-chain data (i.e., DNA fingerprinting). This off-chain data is used for the traceability of products (i.e., products’ origin) along the supply chain with a real-world livestock use case. To this end, we present a livestock blockchain oracle (LBO) as a service to mitigate the accessibility issue and automate the process of verifying purchasable products for livestock DNA fingerprinting verification. We have conducted an evaluation study using real-world livestock data from third-party service providers. Results based on the livestock product information and registered DNA service providers show that LBO is a reliable and responsive decentralized oracle blockchain for verification.

Edge computing has been getting a momentum with ever-increasing data at the edge of the network. In particular, huge amounts of video data and their real-time processing requirements have been increasingly hindering the traditional cloud computing approach due to high bandwidth consumption and high latency. Edge computing in essence aims to overcome this hindrance by processing most video data making use of edge servers, such as small-scale on-premises server clusters, server-grade computing resources at mobile base stations and even mobile devices like smartphones and tablets; hence, the term edge-based video analytics. However, the actual realization of such analytics requires more than the simple, collective use of edge servers. In this paper, we survey state-of-the-art works on edge-based video analytics with respect to applications, architectures, techniques, resource management, security and privacy. We provide a comprehensive and detailed review on what works, what doesn't work and why. These findings give insights and suggestions for next generation edge-based video analytics. We also identify open issues and research directions.

While the real-time analysis of dash cam video is of great practical importance for improving road safety, commercial dash cams lack the resources necessary to perform such video analytics. It is impractical to use clouds for this due to high latency and high bandwidth consumption. In this paper, we present eDashA, the first edge-based system that demonstrates the potential of near real-time video analytics using a network of mobile devices, on the move. In particular, it simultaneously processes videos produced by two dash cams of different angles (outward facing and inward facing dash cams) with one or more mobile devices on the move. Further, we devise several optimization techniques and incorporated them into eDashA. These techniques are simultaneous download and analysis, scheduling, segmentation and early stopping. We have implemented eDashA as an Android app and evaluated it using two dash cams and several heterogeneous smartphones. Experiment results show the feasibility of real-time video analytics on the move.

This paper presents an IRB-approved human study to capture data to build models for human frustration prediction of computer users. First, an application was developed that ran in the user’s computer/laptop/VM with Linux 20.04. Then, the application collected a variety of data from their computers, including: mouse clicks, movements and scrolls; the pattern of keyboard keys clicks; user audio features; and head movements through the user video; System-wide information such as computation, memory usage, network bandwidth, and input/output bandwidth of the running applications in the computer and user frustrations. Finally, the application sent the data to the cloud. After two weeks of data collection, supervised and semi-supervised models were developed offline to predict user frustration with the computer using the collected data. A semi-supervised model using a generative adversarial network (GAN) resulted in the highest accuracy of 90%.

A blockchain is a form of distributed ledger technology where transactions as data state changes are permanently recorded securely and transparently without the need for third parties. Besides, introducing smart contracts to the blockchain has added programmability, revolutionizing the software ecosystem toward decentralized applications. Although promising, the usability of smart contracts is primarily limited to on-chain data without access to the external systems (i.e., off-chain) where real-world data and events reside. This connectability to off-chain data for smart contracts and blockchain is an open practical problem referred to as the “oracle problem” and is defined as how real-world data can be transferred into/from the blockchain. Hence, Blockchain oracles are introduced and implemented in the form of application programming interfaces connecting the real world to the blockchain for mitigating such a limitation. This article studies and analyzes how blockchain oracles provide final feedback (i.e., outcome) to smart contracts and survey blockchain oracle technologies and mechanisms regarding data integrity and correctness. Since the existing solutions are extensive in terms of characteristics and usage, we investigate their structure and principles by classifying the blockchain oracle implementation techniques into two major groups voting-based strategies and reputation-based ones. The former mainly relies on participants’ stakes for outcome finalization, while the latter considers reputation and performance metrics in conjunction with authenticity-proof mechanisms for data correctness and integrity. We present the result of this classification with a thorough discussion of the state of the art and provide the remaining challenges and future research directions in the end.