Earth observation (EO) problems require processing large volumes of data, such as those involved in environmental studies using diverse spatiotemporal variables. In this context, the computing continuum offers a promising model to support EO tasks. It enables low-latency data processing near ground stations and broad data sharing through the cloud. Nevertheless, designing systems in the computing continuum introduces challenges. One is the coordination of distributed entities across different computational environments (e.g., the edge, the fog, or the cloud). Another is managing data across heterogeneous infrastructures. This paper presents OrbiFaaS, a serverless-based method for building Continuum Earth Observation Systems (EOS) using an orbital model. OrbiFaaS organizes a continuum EOS as a system of multiple orbits, each representing a different computational environment arranged around a core of data sources. Orbits near the core correspond to low-latency environments (e.g., the edge), while those farther away exhibit higher latencies (e.g., the cloud). Each orbit contains satellites, which represent service provider infrastructures. Organizations can use these satellites to deploy multiple data management services, such as containerized microservices or functions. Satellites are interconnected using data containers that establish communication channels based on available filesystem, memory, and network resources. We evaluate OrbiFaaS in a case study involving the processing of satellite images across multiple environments. Our experimental results demonstrate the feasibility and efficiency of OrbiFaaS for constructing Earth observation systems.

Privacy-Preserving Data Mining (PPDM) methods prevent unauthorized data disclosure during data analysis tasks executed by untrusted third parties, as in Data Mining as a Service (DMaaS) scenarios. However, PPDM models still present usability, performance, security, and practicality issues. This paper presents an elastic cloud-based model for efficient and flexible PPDM as a Service (PPDMaaS) within the Big Data context. The model transparently couples PPDMs with cloud data management. It is based on a stacked architecture that incorporates parallel and distributed processing patterns at design time to efficiently process large-scale volumes of data and create concurrent PPDM processing streams. A prototype of the elastic cloud model was created to validate and evaluate its usability and efficiency under different cryptography-based PPDMs, which supported at least a 128-bit equivalent security level for the most popular data mining tasks: clustering and classification. Validation tests were done using 16 datasets from the UCI repository. In terms of performance, evaluation was done using 50 artificial datasets that resembled a Big Data scenario. The obtained results revealed the efficiency and suitability of the proposed elastic cloud model to enable PPDMaaS. Thus, it guarantees data privacy and reduces processing times mainly induced by homomorphic encryption (one order of magnitude) without affecting the accuracy of obtained data mining models. The novelty of this work is in its elastic and modular design, which enables seamless integration of PPDM methods into a scalable service architecture. The platform provides customizable, efficient, and privacy-preserving data mining capabilities, addressing key limitations of previous approaches and making it suitable for deployment in real-world DMaaS scenarios.

Context: Organizations increasingly rely on artificial intelligence (AI) and machine learning (ML) to process data, automate tasks, and enhance decision-making. At the same time, the computing continuum enables AI and ML to be deployed closer to data sources, thereby reducing system latency and response time. Objective: Managing applications across this distributed environment is challenging due to the need for manual deployment, integration, and compliance with non-functional requirements (NFRs) such as security and fault tolerance. Therefore, there is a need for frameworks that automate the deployment and execution of computing continuum systems while integrating both functional and non-functional requirements. Method: This paper presents Nez , a design-driven skeleton model for building continuum AI and analytics systems. Nez construction model automatically and transparently integrates AI/ML applications with non-functional components to create continuum systems that are deployed dynamically across multiple distributed infrastructures. Results: We conducted case studies on the processing of medical imagery and satellite imagery to provide automatic and continuous support for decision-makers. Nez has already been deployed at the Mexican hospital, Instituto Nacional de Rehabilitación Luis Gerardo Ibarra Ibarra , to create an AI-based data flow supporting bone cancer diagnosis. The evaluation shows that Nez outperforms state-of-the-art tools such as Nextflow, Makeflow, and Parsl, achieving improvements in response time of 28.46%, 17.46%, and 23.54%, respectively. Conclusion: Nez efficiently transforms organizational data flow designs into continuum computing services. This enables organizations to construct continuum AI-based and analytical systems that account for both functional and non-functional requirements.

Data distribution across different facilities offers benefits such as enhanced resource utilization, increased resilience through replication, and improved performance by processing data near its source. However, managing such data is challenging due to heterogeneous access protocols, disparate authentication models, and the lack of a unified coordination framework. This paper presents DynoStore, a system that manages data across heterogeneous storage systems. At the core of DynoStore are data containers, an abstraction that provides standardized interfaces for seamless data management, irrespective of the underlying storage systems. Multiple data container connections create a cohesive wide-area storage network, ensuring resilience using erasure coding policies. Furthermore, a load-balancing algorithm ensures equitable and efficient utilization of storage resources. We evaluate DynoStore using benchmarks and realworld case studies, including the management of medical and satellite data across geographically distributed environments. Our results demonstrate a 10 % performance improvement compared to centralized cloud-hosted systems while maintaining competitive performance with state-of-the-art solutions such as Redis and IPFS. DynoStore also exhibits superior fault tolerance, withstanding more failures than traditional systems.

Computing continuum systems are emerging as a solution for organizations to process data across diverse infrastructures, reducing latency compared to traditional cloud computing. However, managing I/O operations in such distributed and heterogeneous environments remains an open research challenge. In this paper, we present A-Flow, a model for constructing I/O systems to manage data exchange in computing continuum environments. These systems are built around data distribution patterns defined by structures called abstract communication channels (ACCs). ACCs are established between processing stages using memory, file system, and network resource connections. To prevent resource overload during execution, A-Flow automatically selects the appropriate communication channel based on user-defined criteria such as throughput or resource utilization. We implemented this model in a prototype and evaluated it through a case study focused on managing medical data in HDF5 format across different environments. The evaluation revealed that A-Flow’s ACCs can be integrated into existing stage-based systems found in the state of the art. The results highlight the efficiency and effectiveness of A-Flow in enabling dataflows across heterogeneous infrastructures, addressing key challenges in computing continuum.

Modern object stores expose only coarse, static configurations (replication factor, ACLs, bucket lifecycles) and must be over-provisioned to absorb demand spikes or new security requirements. MictlanX upgrades the storage layer itself with two code-defined programming models: a Responsive-Deployment Model, that lets operators declare elastic regions in YAML/Python, and an Adaptive Data-Placement Model, that drives per-object filters and dynamic replication. A 16-node prototype sustains 20 MB/s under a 0.01s burst while keeping 90% of requests below 0.1s---up to 4× faster than fixed-replica modes---and beats MinIO, Google Drive, and Dropbox by up to 23% throughput with sub-second latencies.

In the era of advanced meteorological data platforms such as Copernicus and Climate Data Store, the frontier of weather forecasting has evolved. The primary challenge is no longer the acquisition of accurate and high-resolution data, but rather the effective integration and utilization of diverse observational datasets to enhance localized weather predictions. Crowd sensed weather data through a network of low-cost, widely distributed weather stations can provide the granular data needed for precise local forecasts. However, this approach introduces challenges such as data integration, consistency, and privacy concerns. Federated Learning (FL) addresses these issues by enabling decentralized data processing while maintaining data privacy.This paper introduces an innovative implementation of a federated learning framework integrated with a cluster of Automated Weather Stations (AWS). The primary objective of this study is to leverage federated learning to enhance the predictive accuracy of the Weather Research and Forecasting (WRF) model by using each weather station not only as a data acquisition point but also as a computational node. This decentralized approach maintains data privacy and security while enabling local training of models, such as Crossformer, Autoformer, and DLinear. These models’ locally trained weights are periodically aggregated on the central server, which updates and redistributes the global model.Based on data collected over two years from two automated weather stations, the experimental results analyze the possibility of improving WRF model predictions for temperature and humidity. This research highlights the potential of Federated Learning in meteorological applications, offering a robust solution for enhancing weather forecast accuracy while ensuring data privacy and efficient resource utilization.

Computing continuum is becoming a solution for organizations to process and analyze data for supporting decision-making processes. In this context, serverless paradigm is arising as a solution to manage continuum computing. However, the management of data storage still represents an obstacle for integrating continuum computing and serverless paradigms into a single solution, as this has to be performed transparently to users through multiple infrastructures. This paper presents StructMesh, a storage framework for serverless continuum systems. This framework is based on a processing plane where functions are managed as patterns, and a data plane based on storage meshes that represent maps of storage resources available in a given infrastructure. The logical interconnection of storage meshes enables organizations to integrate storage resources into a single unified storage service, which creates data exchange channels for continuum processing throughout multiple infrastructures. These meshes include load-balancing and data allocation/location algorithms for transparently and automatically managing the inputs/outputs of functions throughout these channels, as well as non-functional requirement schemes for organizations to manage sensitive data. We developed a framework prototype that harmonizes processing serverless functions with storage functions for building serverless pipeline services. A case study was conducted by using these services for processing meteorological and earth observation data throughout multiple infrastructures. The evaluation revealed the efficiency of StructMesh when managing data through fog and cloud infrastructures. It also showed the feasibility of StructMesh to create and enable continuum data exchange channels for serverless pipelines.

Abstract This article presents the design and implementation of a reliable computing virtual container‐based model with integrity verification for data processing strategies named the reliability and integrity verification (RIV) scheme. It has been integrated into a system construction model as well as existing workflow engines (e.g., Kulla and Makeflow) for composing in‐memory systems. In the RIV scheme, the reliability (R) component is in charge of providing an implicit fault tolerance mechanism for the processes of data acquisition and storage that take place in a data processing system. The integrity verification (IV) component is in charge of ensuring that data transmitted/received between two processing stages are correct and are not modified during the transmission process. To show the feasibility of using the RIV scheme, real‐world applications were created by using different distributed and parallel systems to solve use cases of satellite and medical imagery processing. This evaluation revealed encouraging results as some solutions that assumed the cost (overhead) of using the RIV scheme, for example, Kulla (the Kulla‐RIV solution), achieve better response times than others without the RIV scheme (e.g., Makeflow) that remain exposed to the risks caused by to the lack of RIV strategies.

MeshStore, a fault-tolerant serverless storage model for edge-fog-cloud continuum systems, enables organizations to integrate distributed heterogeneous storage resources into a single unified storage service for the sharing of data through serverless functions deployed on edge-fog-cloud environments to create continuum dataflows. This unified service automatically and transparently manages the input/output data of serverless functions by coupling storage structures, including load-balancing and data allocation/location algorithms. Organizations also can add non-functional requirement properties (e.g., either reliability or security) to the storage structures when managing sensitive data.

This paper presents a continuous delivery/continuous verifiability (CD/CV) framework for IoT dataflows in edge-fog-cloud. In this framework a CD model based on extraction, transformation, and load (ETL) mechanism as well as a directed acyclic graph (DAG) construction, enable end-users to create efficient schemes for the continuous verification and validation of the execution of applications in edge-fog-cloud infrastructures. This framework also provides tools for continuous verification and validation (CV) of predefined execution sequences and the integrity of digital assets using blockchain. CV model converts ETL and DAG into business model, smart contracts in a private blockchain for the automatic and transparent registration of transactions performed by each application in workflows/pipelines created by CD model without altering applications nor edge-fog-cloud workflows. This framework ensures that IoT dataflow delivers verifiable information for organizations to conduct critical decision-making processes with certainty. A containerized parallelism approach solves portability issues and reduces/compensates the overhead produced by CD/CV operations. The talk will also present evaluation results of the CD/CV framework based on a case study where user mobility information is used to identify interest points, patterns, and maps. The experimental evaluation results show the feasibility of CD/CV to register transactions performed in IoT dataflows through edge-fog-cloud in a private blockchain network.

Summary Cloud‐based services have proved useful in several research fields, such as engineering, health science, and astrophysics, to mention a few examples. The computational environmental science community developed a strong need for cloud facilities to store, process, and manage data from observations and numerical models for simulations and forecasts. Weather forecast models and global sensor networks deal with multidimensional geo‐referenced data∖sets. However, environmental data consumer applications usually require a relatively small amount of multidimensional input data slice to analyze a specific area or time interval. Hence, reducing data dimension for information retrieval is mandatory. This paper presents a twofold solution: a technique to load and retrieve the sliced multidimensional data set on different cloud services such as Amazon Web Service (AWS), Google Cloud Platform, and Microsoft Azure. The experimental results performed on these cloud services highlight that the proposed method can significantly speed up the process of loading and retrieving the data slices compared to working with the entire data set in bulk or OPeNDAP server.

This paper presents a continuous delivery/continuous verifiability ( CD/CV ) method for IoT dataflows in edge–fog–cloud. A CD model based on extraction, transformation, and load (ETL) mechanism as well as a directed acyclic graph ( DAG ) construction, enable end-users to create efficient schemes for the continuous verification and validation of the execution of applications in edge–fog–cloud infrastructures. This scheme also verifies and validates established execution sequences and the integrity of digital assets . CV model converts ETL and DAG into business model, smart contracts in a private blockchain for the automatic and transparent registration of transactions performed by each application in workflows/pipelines created by CD model without altering applications nor edge–fog–cloud workflows. This model ensures that IoT dataflows delivers verifiable information for organizations to conduct critical decision-making processes with certainty. A containerized parallelism model solves portability issues and reduces/compensates the overhead produced by CD/CV operations. We developed and implemented a prototype to create CD/CV schemes, which were evaluated in a case study where user mobility information is used to identify interest points, patterns, and maps. The experimental evaluation revealed the efficiency of CD/CV to register the transactions performed in IoT dataflows through edge–fog–cloud in a private blockchain network in comparison with state-of-art solutions.

This paper presents the design, development, and evaluation of PuzzleMesh, an agnostic service mesh composition model to process large volumes of data in edge-fog-cloud environments. This model is based on a puzzle metaphor where pieces, puzzles, and metapuzzles represent self-contained autonomous and reusable software artifacts encapsulated into containers and published as microservices. Apiecerepresents the integration of apps with I/O interfaces (loops/sockets), parallel processing, and management software. Apuzzlerepresents a processing structure (e.g., workflows) built coupling pieces through loops and sockets. Puzzles integrate structures with a microservice architecture, implicit continuous dataflows, and transparent data exchange management software. Ametapuzzlerepresents a recursive assemble of puzzles. A mesh represents a pool of pieces, puzzles, and metapuzzles available for designers to choose artifacts to build services. A prototype developed using PuzzleMesh model was evaluated through case studies about the automatic construction of processing services for the acquisition, pre-processing, manufacturing, preserving, and visualizing of satellite imagery. A qualitative comparison revealed that PuzzleMesh provides a flexible way to build reusable and portable services and to improve the usability of the services. The case study also revealed that PuzzleMesh yielded better performance results than other state-of-the-art tools.

The systems that meet non-functional requirements (NFRs) are key for e-health services to face up events such as service outages and violations of confidentiality. How-ever, traditional NFR systems produce overhead in execution time, which could affect critical decision-making processes. This paper presents a dynamic parallel pattern construction model to design and create efficient NFR self-adaptable sys-tems. The construction of patterns is performed in two design phases: in the first one, the designers build NFR systems by creating pipelines including as many applications as required to meet the NFRs established by healthcare organizations. In the second phase, a pipeline is converted into a worker that auto-matically is added to a dynamic pattern. In a dynamic pattern, the workers can be cloned to be executed by different parallel patterns (e.g., manager/worker, divide&conquer, etc.) to face changes in the incoming workload during execution time, which converts a worker into a self-adaptable NFR system. A proto-type was implemented to create self-adaptable NFR systems, which were used in a case study to manage spirometry studies, tomography images, and electrocardiograms. The evaluation showed the effectiveness of this dynamic pattern model to create self-adaptable systems when processing multiple types of medical data/contents. The evaluation also revealed that the self-adaptable NFR systems built by dynamic patterns yielded significant performance gain in a direct comparison with the implementation of NFR application pipelines built by a traditional framework called Jenkins.

Cloud storage has been the solution for organizations to manage the exponential growth of data observed over the past few years. However, end-users still suffer from side-effects of cloud service outages, which particularly affect edge-fog-cloud environments. This paper presents SeRSS, a storage mesh architecture to create and operate reliable, configurable, and flexible serverless storage services for heterogeneous infrastructures. A case study was conducted based on-the-fly building of storage services to manage medical imagery. The experimental evaluation revealed the efficiency of SeRSS to manage and store data in a reliable manner in heterogeneous infrastructures.

Content delivery and sharing (CDS) is a popular and cost effective cloud-based service for organizations to deliver/share contents to/with end-users, partners and insider users. This type of service improves the data availability and I/O performance by producing and distributing replicas of shared contents. However, such a technique increases overhead on the storage/network resources. This article introduces a threefold methodology to improve the trade-off between I/O performance and capacity utilization of cloud storage for CDS services. This methodology includes: i) Definition of a classification model for identifying types of users and contents by analyzing their consumption/ demand and sharing patterns, ii) Usage of the classification model for defining content availability and load balancing schemes, and iii) Integration of a dynamic availability scheme into a cloud-based CDS system. Our method was implemented on both a simulator and a real-world CDS service, supporting information sharing operations performed in a cloud storage. An experimental evaluation, conducted in a private cloud through simulation and emulation of workloads, showed the feasibility of this methodology in terms of storage capacity utilization, whereas the real-world implementation revealed the efficiency of applying a classification model to information sharing patterns in terms of I/O performance.

In organizational environments, such as in hospitals, data have to be processed, preserved, and shared with other organizations in a cost-efficient manner. Moreover, organizations have to accomplish different mandatory non-functional requirements imposed by the laws, protocols, and norms of each country. In this context, this paper presents a Federated Content Distribution System to build infrastructure-agnostic health data synchronization services. In this federation, each hospital manages local and federated services based on a pub/sub model. The local services manage users and contents (i.e., medical imagery) inside the hospital, whereas federated services allow the cooperation of different hospitals sharing resources and data. Data preparation schemes were implemented to add non-functional requirements to data. Moreover, data published in the content distribution system are automatically synchronized to all users subscribed to the catalog where the content was published.

A major concern of users of cloud storage services is the lost of control over security, availability and privacy of their files. That is partially addressed by end-to-end encryption techniques. However, most of the solutions currently available offer rigid functionalities that cannot be rapidly integrated into customized tools to meet user's requirements like, for example, file sharing with other users. This paper presents an end-to-end architecture that enables users to build secure and resilient work-flows for storing and sharing files in the cloud. The workflows are configurable structures executed on the user-side that perform processing operations on the files through chained stages such as data compression for capacity overhead reduction, file assurance for ensuring confidentiality when sharing files and information dispersion for storing files in n cloud locations and retrieving them even during outages of m cloud storage providers. The users can set up different workflows depending on their requirements because they can organize the processing units of each stage in either pipeline to improve its performance or stack for improving functionality. The stages and their processing units are connected using I/O communication interfaces which ensure a continuous data flow from the user/organization computers to multiple cloud locations. Based on our architecture, we developed a prototype for a private cloud infrastructure. The experimental evaluation revealed the feasibility of enabling flexible file sharing and storage user-defined workflows in terms of performance.

This paper presents, in the form of a case study, the design, implementation and performance evaluation of Phoenix: a prototype of fault-tolerant distributed storage web service, based on URLs. Phoenix includes Web Storage Middleware (WSM) that re-uses the URLs of the files to accommodate data/redundancy load dynamically as well as a set of redundancy strategies. Strategies include simple file-replication, error-coding techniques and a new hybrid adaptive technique. Experimental evidence on performance shows that distributing the users to several web servers compensates the overhead produced both when generating and distributing data redundancy. We also found during web server failures, users do not perceive but a lengthening on the system response times. As it could be expected, peak users requests and network congestion are the most important factors affecting the performance of storage system.