Herodotos Herodotou

Background With the increasing volume of data collected for advanced analytical and AI applications, data storage remains a significant challenge. Despite advancements in storage technologies, the cost of maintaining vast datasets continues to grow. Compression techniques have been widely used to address this issue, but existing systems primarily rely on a single, typically lossless method, which limits adaptability to varying data characteristics. Methods This paper introduces COMPASS, a multiple compression approach that applies different compression techniques to distinct subsets of data within a database. COMPASS partitions relational data into rows or columns and selects the most suitable compression scheme for each column or column group. Two versions of COMPASS are proposed: (i) COMPASS-D, which utilizes K-Means clustering based on data values; and (ii) COMPASS-E, which employs K-Means clustering based on column entropy to group similar columns efficiently. The effectiveness of COMPASS is evaluated using the Envmon dataset, a real-world environmental monitoring database, and compared against monolithic compression methods. Results Experimental results demonstrate that COMPASS significantly reduces disk space usage compared to traditional compression techniques. COMPASS-E achieves superior performance in terms of compression time and proximity to the optimal compression ratio, outperforming COMPASS-D. In worst-case scenarios, COMPASS methods offer 22% more savings compared to baseline techniques, with best-case savings reaching 56% (~2× improvement). Conclusion The proposed COMPASS framework offers a flexible and adaptive approach to database compression by leveraging multiple schemes tailored to different data subsets. This results in improved storage efficiency and reduced computational overhead. Future work will explore additional data characteristics and clustering methods to further enhance COMPASS’s adaptability and efficiency.

This study examines the operational resilience of Limassol Port, Cyprus’s primary maritime hub, amid disruptions caused mainly by the COVID-19 pandemic. Utilizing high-resolution port-level data from 2018 to 2025, we evaluate performance across five key performance indicators: port calls, anchorage utilization, berth utilization, waiting times, and arrival punctuality. The analysis adopts a longitudinal approach, spanning pre-pandemic, peak-pandemic, post-pandemic, and recent phases, while differentiating impacts across vessel categories. Unlike many regional ports, Limassol’s cruise sector exhibited unique counter-cyclical growth, with calls doubling during the pandemic as the port transitioned into a strategic safe haven and repositioning base. This surge normalized over the 2024–2025 period as temporary operational disruptions resolved. Conversely, container and Ro-Ro (roll-on/roll-off) segments demonstrated robust stability, achieving rapid post-pandemic normalization, while bulk and tanker operations exhibited higher volatility linked to shifting commodity demands. These findings, validated through one-way analysis of variance (ANOVA) and Cohen’s d effect sizes, underscore the adaptive capacity of mid-sized Mediterranean hubs. The study concludes that operational flexibility, coupled with enhanced digital coordination and strategic capacity planning, is essential for maintaining the resilience of short sea shipping networks during global crises.

Modern data storage systems contain numerous media devices (e.g., DRAM, NVRAM, SSDs, and HDDs) organized into multi-level caches and storage tiers to optimize the performance-to-cost ratio. Traditional caching policies for eviction, admission, and prefetching, as well as tiering policies for data placement and migration, offer simplicity but struggle to adapt to the dynamic and complex access patterns observed in modern data-intensive applications. As a result, machine learning (ML)-based policies have recently gained popularity due to their ability to make predictions and proactively optimize caching and tiering operations. These ML techniques not only increase cache hit rates and reduce latency but also provide scalable, cost-effective solutions that intelligently place data across different storage media. This manuscript reviews the state-of-the-art ML-based caching and tiering approaches, examining their theoretical foundations and practical implementations. It also presents the most common features for each policy type, the most popular baselines for comparison, and the typical evaluation metrics. Finally, it discusses emerging trends and outlines potential directions for future research in data storage systems.

Recently, the need for unified orchestration frameworks that can manage extremely heterogeneous, distributed, and resource-constrained environments has emerged due to the rapid development of cloud, edge, and IoT computing. Kubernetes and other traditional cloud-native orchestration systems are not built to facilitate autonomous, decentralized decision-making across the computing continuum or to seamlessly integrate non-container-native devices. This paper presents the Distributed Adaptive Cloud Continuum Architecture (DACCA), a Kubernetes-native architecture that extends orchestration beyond the data center to encompass edge and Internet of Things infrastructures. Decentralized self-awareness and swarm formation are supported for adaptive and resilient operation, a resource and application abstraction layer is established for uniform resource representation, and a Distributed and Adaptive Resource Optimization (DARO) framework based on multi-agent reinforcement learning is integrated for intelligent scheduling in the proposed architecture. Verifiable identity, access control, and tamper-proof data exchange across heterogeneous domains are further ensured by a zero-trust security framework based on distributed ledger technology. When combined, these elements enable increasingly autonomous workload orchestration, trading centralized control for adaptive, decentralized operation with enhanced interoperability, scalability, and trust. Thus, the proposed architecture enables self-managing and context-aware orchestration systems that support next-generation AI-driven distributed applications across the entire computing continuum.

Data prefetching is essential for modern file storage systems operating in large-scale cloud and data-intensive environments, where high performance increasingly depends on intelligent, adaptive mechanisms. Traditional rule-based methods and recently proposed machine learning-based techniques often struggle to cope with the complex and rapidly evolving data access patterns characteristic of big-data workloads. In this paper, we introduce an online, streaming machine learning (SML) approach for predictive data prefetching that retrieves useful data into the cache ahead of time. We present a novel online training framework that extracts features in real time and continuously updates streaming ML models to learn and adapt from large and dynamic access streams. Building on this framework, we design new SML-driven prefetching algorithms that decide when, how, and what data to prefetch into the cache with minimal overhead. Extensive experiments using production traces from Huawei Technologies Inc. and Google workloads from the SNIA IOTTA repository demonstrate that our intelligent policies consistently deliver the highest byte hits among competing approaches, achieving 97% prefetch byte precision and reducing data access latency by up to 2.8 times. These results show that streaming ML can deliver immediate performance gains and offers a scalable foundation for future adaptive storage systems.

Recommendation systems are popular information systems that help consumers manage information overload. Whilst personality has been recognised as an important factor influencing consumers’ choice, it has not yet been fully exploited in recommendation systems. This study proposes a restaurant recommendation approach that integrates customer personality traits, opinions and preferences, extracted either directly from online review platforms or derived from electronic word of mouth (eWOM) text using information extraction techniques. The proposed method leverages the concept of venue personality grounded in personality–brand congruence theory, which posits that customers are more satisfied with brands whose personalities align with their own. A novel model is introduced that combines fine-tuned BERT embeddings with linguistic features to infer users’ personality traits from the text of their reviews. Customers’ preferences are identified using a custom named-entity recogniser, while their opinions are extracted through structural topic modelling. The overall framework integrates neural collaborative filtering (NCF) features with both directly observed and derived information from eWOM to train an extreme gradient boosting (XGBoost) regression model. The proposed approach is compared to baseline collaborative filtering methods and state-of-the-art neural network techniques commonly used in industry. Results across multiple performance metrics demonstrate that incorporating personality, preferences and opinions significantly improves recommendation performance.

Recently, the need for unified orchestration frameworks that can manage extremely heterogeneous, distributed, and resource-constrained environments has arisen due to the rapid development of cloud, edge, and IoT computing. Kubernetes and other traditional cloud-native orchestration systems are not built to facilitate autonomous, decentralized decision-making across the computing continuum or to seamlessly integrate non-container-native devices. This paper presents the Distributed Adaptive Cloud Continuum Architecture (DACCA), a Kubernetes-native architecture that extends orchestration beyond the data center to encompass edge and Internet of Things infrastructures. Decentralized self-awareness and swarm formation are supported for adaptive and resilient operation, a resource and application abstraction layer is established for uniform resource representation, and a Distributed and Adaptive Resource Optimization (DARO) framework based on multi-agent reinforcement learning is integrated for intelligent scheduling in the proposed architecture. Verifiable identity, access control, and tamper-proof data exchange across heterogeneous domains are further guaranteed by a distributed-ledger-technology-based zero-trust security framework. When combined, these elements enable completely autonomous workload orchestration with enhanced interoperability, scalability, and trust. Thus, the proposed architecture enables self-managing and context-aware orchestration systems that support next-generation AI-driven distributed applications across the entire computing continuum.