Experimentation on distributed, heterogeneous computing environments—from edge devices to large-scale cloud platforms—demands orchestration technologies that are both flexible and extensible. Kiso is an open-source framework designed to provision resources and manage complex scientific workflows across the edge-to-cloud continuum. Its architecture unifies infrastructure provisioning, experiment configuration, and reproducible execution, enabling researchers to compose and monitor experiments that span geographically dispersed sites and variable network conditions. Although Kiso was conceived for workflow management—coordinating data-intensive tasks and ensuring reproducibility across dynamic infrastructures—its modular design makes it equally promising for providing reproducible environments for deploying and studying emerging agentic frameworks, where autonomous AI agents require consistent resource provisioning, cross-site communication, and result collection. We describe Kiso’s core capabilities for resource orchestration, experiment lifecycle management, and integration with containerized services, and we outline how these capabilities can support distributed multi-agent systems. In particular, we discuss how its declarative provisioning, extensible task abstractions, and built-in monitoring and output collection provide a natural foundation for experiments in which reasoning agents plan, negotiate, and adapt in real time. This study situates Kiso at the intersection of scientific workflow management and complex, agent-based computing, highlighting its potential to accelerate research on adaptive, self-organizing cyber-physical systems—an emerging frontier in complex systems science.

The term scientific workflow has evolved over the last two decades to encompass a broad range of compositions of interdependent compute tasks and data movements. It has also become an umbrella term for processing in modern scientific applications. Today, many scientific applications can be considered as workflows made of multiple dependent steps, and hundreds of workflow management systems (WMSs) have been developed to manage and run these workflows. However, no turnkey solution has emerged to address the diversity of scientific processes and the infrastructure on which they are implemented. Instead, new research problems requiring the execution of scientific workflows with some novel feature often lead to the development of an entirely new WMS. A direct consequence is that many existing WMSs share some salient features, offer similar functionalities, and can manage the same categories of workflows but also have some distinct capabilities. This situation makes researchers who develop workflows face the complex question of selecting a WMS. This selection can be driven by technical considerations, to find the system that is the most appropriate for their application and for the resources available to them, or other factors such as reputation, adoption, strong community support, or long-term sustainability. To address this problem, a group of WMS developers and practitioners joined their efforts to produce a community-based terminology of WMSs. This paper summarizes their findings and introduces this new terminology to characterize WMSs. This terminology is composed of fives axes: workflow characteristics, composition, orchestration, data management, and metadata capture. Each axis comprises several concepts that capture the prominent features of WMSs. Based on this terminology, this paper also presents a classification of 23 existing WMSs according to the proposed axes and terms.

We present a unified integer-only kernel for hyperbolic functions tanh(x) and sigmoid(x), designed for embedded RISC-V platforms without floating-point units. The kernel combines LUT anchoring, CORDIC microrotations, and linear correction in Q20 arithmetic, producing outputs directly in Q1.16 format suitable for ML inference. Exhaustive evaluation across the entire 17-bit input space demonstrates deterministic accuracy better than 1 ULP in Q1.16, with maximum absolute errors below 1.5 × 10^-5 . Compared to software-emulated floating-point implementations, our approach achieves significant speedup while reducing hardware complexity by unifying multiple activation functions in a single IP block. This makes the method highly relevant for efficient deployment of neural networks on resource-constrained RISC-V microcontrollers.

Distributed scientific workflows are increasingly executed across heterogeneous and geo-distributed computing environments, where centralized workload orchestration becomes a scalability and resilience bottleneck. This paper presents SWARM+, a decentralized workload management system that coordinates workload placement through hierarchical multi-agent consensus, reducing coordination overhead and dramatically improving scalability, while tolerating failures and dynamic membership changes. SWARM+ enables data-aware scheduling policies that incorporate resource availability, data transfer node (DTN) connectivity, and data locality into workload placement decisions. We evaluate SWARM+ on the distributed FABRIC testbed using heterogeneous scientific workloads derived from production workflow traces obtained from the Pegasus Workflow Management System (WMS). Experimental results show that SWARM+ scales coordination to 990 distributed agents with sub-second per-job selection time with 110 agents. SWARM+ demonstrates balanced workload distribution, maintains over 95% job completion under distributed failures with graceful degradation during correlated site outages, tolerates coordinator agent failures gracefully, improves schedule quality by employing data-aware policies, and reduces both selection time and scheduling latency by 97–98% when compared to the prior SWARM system.

Drone fleets equipped with onboard cameras, computer vision, and Deep Neural Network (DNN) models offer a powerful solution for diverse spatio-temporal applications. In the context of wildfire response, drones can play a critical role in monitoring fire spread, assisting firefighters in suppression efforts, and facilitating safe evacuation. In this paper, we introduce AeroResQ, an edge-accelerated Unmanned Aerial Vehicle (UAV) framework for scalable, resilient, and collaborative escape route planning in dynamic wildfire scenarios. AeroResQ introduces service drones (SDs) and coordinator drones (CDs), each with specialized roles. SDs survey fire-affected areas, detect stranded individuals using onboard edge accelerators running fire detection and human pose identification DNN models, and issue assistance requests. CDs, equipped with lightweight data stores such as Apache IoTDB (Database for Internet of Things), dynamically generate optimal ground escape routes and monitor firefighter movements along these routes. AeroResQ employs a collaborative path-planning approach using a weighted A* search algorithm based on terrain constraints, fire spread, and evacuees' positions. AeroResQ's workflows incorporate intelligent load-balancing and resilience strategies: CD failures trigger automated data redistribution across IoTDB replicas, while SD failures prompt spatial re-partitioning of the fire region and reallocation of these partitions to existing SDs. We evaluate AeroResQ on an emulated version of the 2025 Southern California wildfires, analyzing escape route generation time, firefighter coordination efficiency, and system responsiveness. Our results indicate that AeroResQ achieves a median end-to-end latency of <= 500 ms for different fleet configurations, fire regions, and experiment durations of up to 180 minutes, demonstrating its feasibility for real-time deployment in emergency response.

Scientists increasingly rely on sensor-based data; however transforming raw streams into insights across the edge-to-cloud continuum remains difficult due to the breadth of expertise required to coordinate the necessary data and computation flow. This paper introduces a pattern-based, AI-assisted methodology for rapid development of sensor-driven applications. Using Pegasus workflows executing on the FABRIC testbed, we demonstrate a 5-step development loop that shifts workflow construction and deployment from code-first to intent-first design. Starting from an existing Orcasound hydrophone workflow as a reusable template, we generate and refine workflows for air quality, earthquake, and soil moisture monitoring applications. We further show how these workflows extend to edge resources-including BlueField-3 DPUs and Raspberry Pis-through configuration and placement rather than workflow redesign. Our evaluation, from the perspective of a novice Pegasus user, shows that AI-assisted pattern reuse compresses multi-stage workflow development to 1-1.5 days per workflow while preserving the rigor and portability of workflow-based execution.

BBRv3, the latest iteration of Google’s BBR congestion control algorithm, has shown significant performance improvements in high-bandwidth networks. However, our analysis reveals that BBRv3’s startup phase can lead to persistent fairness issues, where flows that initially acquire a larger bandwidth share maintain their advantage throughout the connection lifetime. To address this, we propose three enhanced BBRv3 variants that modify the startup behavior. Our evaluation on the FABRIC testbed demonstrates that these variants significantly improve fairness metrics, with BBRv3e1 achieving up to 15% improvement in Jain’s fairness index while preserving over 95% link utilization across diverse network scenarios. These improvements enable more equitable resource allocation in high-bandwidth networks, ensuring that BBRv3 can be deployed with confidence.

The Compute Continuum—spanning IoT, Edge, Cloud, and HPC resources—is reshaping how hyper-distributed applications are designed and orchestrated. Traditional service orchestrators and workload management systems rely on centralized runtimes; however, the emerging paradigm requires decentralized coordination, where autonomous agents cooperate to achieve common goals and dynamically distribute workloads. Consensus algorithms play a crucial role in multi-agent systems (MAS), as they enable agents to reach agreement on how to coordinate and execute functionalities in a cooperative manner. While consensus has previously been applied to distributed job selection, here we extend its use to swarm environments. In this setting, agents autonomously decide which service functionalities (i.e., roles) to execute based on their capabilities and the real-time quality of service (QoS). Functionalities can be elastically activated or terminated as application needs evolve. To support this model, we leverage the COLMENA framework, a programming environment for defining and managing such dynamic services. We apply a greedy consensus-based approach to modern power systems, which are increasingly decentralized due to the large-scale integration of renewable energy sources. Centralized power plants are giving way to distributed, intermittent resources that require decentralized control paradigms. To demonstrate this, we simulate the Northeastern Power Coordinating Council’s (NPCC) 140-bus grid using the ANDES simulator in conjunction with the COLMENA middleware. We deploy this use case across six different sites in the FABRIC testbed, using up to 60 different nodes. Our results show that, under contingency scenarios such as load and generator disconnections, agents self-organize, elect local leaders, and execute optimization algorithms to stabilize grid frequency. Detection and organization times remain below 10s across all experiments, even as the number of agents per area scales from 3 to 10. Stability is restored within approximately 27s and 40s for the respective cases. Resource overhead is minimal, with CPU and memory usage remaining below 7.5% and 2%, respectively. Experiment automation and reproducibility are ensured through Kiso. These findings indicate that role-based programming models complement traditional workflows and that consensus-driven coordination can effectively decentralize decision-making in swarm environments. This approach represents a step toward enabling resilient, decentralized power systems.

Scientists increasingly rely on sensor-based data, yet transforming raw streams into insights across the edge-to-cloud continuum remains difficult. Provisioning heterogeneous infrastructure and managing execution on emerging platforms like Data Processing Units typically requires cross-domain expertise, creating significant barriers to rapid prototyping. This paper introduces an experience-driven methodology for the rapid development of sensor-driven applications. By combining pattern-based workflow engineering with AI-assisted development-implemented via Pegasus on the FABRIC testbed - we utilize an existing Orcasound hydrophone workflow as a reusable template. We introduce a pattern-based engineering methodology to generate and refine workflows for air quality, earthquake, and soil moisture monitoring. Furthermore, we show how these abstract structures are extended to edge resources through modular configuration and placement. Our evaluation focuses on user productivity and practical lessons rather than peak performance. Through these case studies, we illustrate how AI-assisted, pattern-based development lowers the entry barrier for non-experts and enables iterative exploration of sensor-driven applications across distributed infrastructures.

Federated learning over medical images is a demanding workflow application. Each round fans out across parallel client jobs and converges on an aggregation step that feeds the next round. At scale this yields 101 sub-workflows and 2,679 jobs on GPUs at four sites, which takes an expert months to build, mostly on workflow mechanics rather than science. We ask how far AI assistance can automate such workflows. An LLM agent, grounded in a released plugin of Pegasus-specific skills, first produces a reviewable specification of checkable constraints and acceptance criteria, then generates the executable workflow. A validation loop repairs runtime failures, checks code against those constraints, and regenerates the implementation from the specification alone. We evaluate three LLM agents, report end-to-end runs on the FABRIC testbed, and show how conformance checking against the specification caught three silent errors that failure-driven debugging missed, including one that trained 1,700 jobs on random tensors.

Distributed computing systems require accurate network simulation tools to optimize data transfer and resource allocation across geographically distributed facilities. We extend DISTRI, a discrete-event simulator for multi-facility distributed computing, with a comprehensive Transmission Control Protocol (TCP) stack supporting multiple congestion control algorithms (CCAs) and network topologies. This unified TCP implementation enables realistic simulation of both inter-facility wide-area and intra-facility local network communications. We validate DISTRI’s TCP stack accuracy by comparing inter-facility scenarios against real-world experiments on the FABRIC testbed. Using a dumbbell topology with competing data transfers, we analyze TCP Reno’s behavior through congestion window (CWND) evolution, round-trip time (RTT), throughput, and packet loss patterns. Results demonstrate that DISTRI accurately captures essential TCP dynamics, with behavioral trends closely matching real-world observations. This validation establishes DISTRI as a reliable, cost-effective tool for developing and testing network optimization algorithms in distributed computing environments. Beyond qualitative validation, we outline directions for incorporating more detailed quantitative error analysis between simulated and measured TCP traces as part of future work.

CORDIC-based architectures remain a key building block for implementing elementary transcendental functions on resource-constrained FPGAs. However, classical CORDIC suffers from long latencies, irregular iteration patterns in hyperbolic mode, and non-trivial scale-factor compensation. This work builds on a low-latency recoding CORDIC for trigonometric functions based on look up tables anchoring, angle recoding and linear interpolation. It extends that framework to support the hyperbolic mode and the exponential function exp(x), and validates it across two FPGA families (Intel/Altera Cyclone IV and AMD/Xilinx Spartan-7). The proposed recoded hyperbolic kernel uses a monotonic shift schedule, eliminates runtime direction computation, and leverages a bit-accurate fixed-point model to derive all constants and interpolation coefficients. Experimental evaluation with 16-bit fixed-point operands compares the proposed design with vendor CORDIC and exp IP cores. For exp(x), the framework achieves a latency reduction of about 36% on the Spartan-7 device relative to the AMD IP core, while using a similar number of LUTs; on Cyclone IV it reduces logic consumption by around 10% and nearly halves the latency with respect to the Intel exp IP core. For sine and cosine, the recoded architecture requires substantially fewer LUTs than vendor CORDIC blocks while preserving numerical accuracy. Overall, the extended framework provides a unified, portable and low-latency solution for trigonometric and exponential computations on FPGAs, suitable for edge and embedded environments.

The landscape of high-performance computing (HPC) systems keeps expanding with new technologies and increased system complexity. In hardware, this trend is driven by technological innovations providing increasing computing power and storage capabilities while taming cost and power metrics. For software, we are witnessing a steady increase in the complexity of algorithms, in data sizes, and in processing requirements. Successfully handling this expansion and maintaining performance and efficiency at adequate levels requires the existing and emerging hardware, middleware, and application aspects and concerns to be re-examined and adapted to the evolving and new paradigms. The goal of this Special Issue is to provide insights into recent advances in innovative technologies and models, algorithms, and software tools that facilitate efficient and convenient use of emerging HPC platforms and enable large-scale applications, including techniques based on artificial intelligence and machine learning. Particular attention is paid to the future of computing beyond Moore’s Law.

Quantum computing exhibits the unique capability to natively and efficiently encode various natural phenomena, promising theoretical speedups of several orders of magnitude. However, not all computational tasks can be efficiently executed on quantum machines, giving rise to hybrid systems, where some portions of an application run on classical machines, while others utilize quantum resources. Efforts to identify quantum candidate code fragments that can meaningfully execute on quantum machines primarily rely on static code analysis. Yet, the state-of-the-art in static code analysis for quantum candidates remains in its infancy, with limited applicability to specific frameworks and languages, and a lack of generalizability. Existing methods often involve a trial-and-error approach, relying on the intuition and expertise of computer scientists, resulting in varying identification durations ranging from minutes to days for a single application. This paper aims to systematically formalize the process of identifying quantum candidates and their proper encoding within classical programs. Our work addresses the critical initial step in the development of automated reasoning techniques for code-to-code translation, laying the foundation for more efficient quantum software engineering. Particularly, this study investigates a sociotechnical phenomenon where the starting point is not a problem directly solvable with QC, but rather an existing classical program that addresses the problem. In doing so, it underscores the interdisciplinary nature of QC application development, necessitating collaboration between domain experts, computer scientists, and physicists to harness the potential of quantum computing effectively.

Current approaches to resilience for highly distributed, heterogeneous, large-scale scientific workflows are limited. Most existing workflow and resource management systems have a single point of failure and resilience strategies are often static, depend on a centralized control, and require considerable design effort from experts. The increasing scale and complexity of workflows coupled with limited resilience capabilities in centralized systems necessitates a fully decentralized, adaptive resource management approach. This paper addresses a very important slice of the overall problem by leveraging the advances in multi-agent systems (MAS). In particular, we explore the suitability of a MAS consisting of globally distributed agents to perform distributed job selection from a dynamic job pool in a truly decentralized, performant, and resilient manner. We present a novel consensus formulation of the distributed job selection problem. By introducing a cost function encapsulating the requirements and constraints of the job and resource loads, we design a novel, greedy consensus algorithm leveraging the Practical Byzantine Fault Tolerance (PBFT)-based consensus method, allowing agents to collectively select jobs in a resilient manner. We compared our algorithms with other state of the art approaches by deploying them in a network testbed infrastructure to emulate distributed job selection. Our evaluation results demonstrated that our greedy consensus algorithm employing the cost-function and PBFT-based consensus method outperforms the ones using the vanilla PBFT-based consensus method - improving scheduling latency by as much as 63.5 % and reducing resource idle time by as much as 63.8 %, with benefits increasing with higher numbers of agents emulated.

This article is an extended version of our conference paper "Modified CORDIC Algorithm for Givens Rotator" (Poczekajlo et al., 2024) published at the International Conference on Computational Science (ICCS-2024). The CORDIC algorithm is an iterative method of computing trigonometric functions and rotating vectors without using complex calculations. This paper presents two modified CORDIC algorithms for implementing a Givens rotator on FPGA, improving upon classic CORDIC methods. The first approach introduces a selective iteration scheme with an optimized scaling factor, while the second, not published in the original ICCS-2024 paper, leverages a scaling-free methodology for improved precision. Implemented on an Altera Cyclone V FPGA, these algorithms demonstrate a 50% accuracy improvement and a 15% reduction in latency compared to standard methods. These findings contribute to enhanced FPGA-based trigonometric computations, particularly benefiting real-time signal processing and numerical linear algebra applications.

Modern scientific workflows process massive amounts of data from diverse instruments and sensors, leveraging geographically distributed, heterogeneous compute and storage resources—from leadership-class systems to edge devices—connected by high-performance networks. The diversity of resources introduces challenges in harnessing their full potential, with resilience issues arising across applications, system software, networks, storage, and hardware. Today, workflow management systems (WMS) coordinate the execution of computation and data management tasks across target resources. However, WMS’s centralized nature makes them vulnerable to faults and scalability issues that may result in failures of entire computational campaigns. This paper introduces a novel agentic framework for workflow management, fully distributing and decentralizing the WMS functions and modeling them as swarm intelligence agents infused with advanced artificial intelligence solutions and traditional distributed computing algorithms that can make coordinated decisions in the presence of failures of the underlying cyberinfrastructure.