In uncertain environments, compliance assurance for business processes faces the intertwined challenges posed by diverse non-functional properties (NFPs) and complex gateways in the models. While solutions like user-defined metrics and constraints empower businesses to act independently, or Markov Decision Process (MDP) enhance statistical algorithm design, they address uncertainties from different views and have not been considered together to tackle a holistic, integrated business and system uncertainties. This paper proposes Process-aware MDP (P-MDP), a framework that unifies the two divergent view to optimize NFPs for business processes operating under uncertainty. We devise a gateway-aware, lazy-evaluation reward mechanism supporting the key stakeholders – business managers – to customize metrics, algorithms, and constraints, and apply them on business processes. Experiments with the WSDREAM benchmark dataset show that P-MDP outperforms the state-of-the-art (SOTA) method constraint-satisfied service composition MDP (CSSC-MDP) at various scales. Moreover, P-MDP demonstrates superior generality and scalability, enabling stakeholders to generate better execution plans for business processes in complex scenarios.

This paper presents a conceptual view on a data platform of a distributed digital twin (DT) architecture for edgecloud software systems, of which entities are provided and built by multiple stakeholders in a supply chain. We focus on exploring data platform approaches for DTs to support security-based assurance development and operation activities in an integrated manner by considering a holistic view of data, AI/ML models and service-based components as part of the edge-cloud software.

Multi-modality machine learning (ML) systems are increasingly deployed in edge environments, such as smart building and autonomous robotic applications. These systems enable analysis of complex subjects by performing ML inferences on multiple data sources provided by Internet of Things (IoT)-enabled devices. However, operating such ML systems for multi-tenant applications presents a runtime orchestration challenge to optimize the execution of multiple inference tasks across modalities. Analysis requests from different tenants often rely on different data sources with distinct quality requirements, resulting in a wide range of conflicting optimization objectives for runtime inference orchestration. The diversity of ML models and their performance variations further complicate the orchestration, particularly in scheduling inference tasks and distributing inference workloads on a heterogeneous edge system. This article addresses these challenges by introducing an adaptive orchestration that manages inference task scheduling for multi-tenant applications in a multi-modality ML system. Our orchestration supports a complex tradeoff between quality refinements and inference capability of inference services while scheduling multiple inference tasks under dynamic time constraints. The orchestration employs an efficient mechanism for selecting instances of inference services to distribute inference workloads across a heterogeneous edge cluster. This mechanism allows the orchestration to leverage cross-modal information to refine inference quality following the tenant objectives. Furthermore, our orchestration considers the performance variation caused by resource contentions, runtime failures, and explainability overheads while distributing workloads. Extensive experiments on two real-world edge applications in a common heterogeneous edge testbed demonstrate that our orchestration can improve inference accuracy by up to 10% and reduce the late response ratio by up to 18% compared to existing approaches within our multi-tenant edge scenarios.

The proliferation of edge data availability alongside advanced federated and distributed machine learning training techniques calls for new developments of machine learning (ML) with distributed and private edge data providers. Most existing works, however, focus on the development and optimization of federated ML communication and aggregation techniques and under-research the quality of training based on the impact of the quality of data and contributions of distributed data sources from such providers to the building of ML models. In this paper, we introduce an Edge marketplAce for DistRibuted AI/ML traiNing (EADRAN), a comprehensive platform for federated learning (FL) with independent edge data providers. The key distinguishable feature of EADRAN is to enable the explainable quality of training (eQoT) approach based on the quality of data and the contributions of provided data to the target-trained ML models. EADRAN offers end-to-end services for both data providers and ML market consumers, enhancing explainability and incentivizing the active participation of edge data providers. EADRAN shows high adaptability via the integration with notable FL frameworks like Flower to allow consumers to choose their preferred FL methods. EADRAN is developed based on a detailed conceptual architecture, guided by novel principles and requirements for federated machine learning marketplaces. We present the detailed design and implementation of EADRAN and conduct extensive experiments to demonstrate the benefits of eQoT-aware training in EADRAN with heterogeneous data scenarios.

Leveraging AI/GenAI and data processing techniques to extract data from technical documents has proliferated due to recent advances in LLM capabilities and open source tools. However, being able to contextualize suitable GenAI/LLMs capabilities coupled with enterprise constraints on cost, data regulation, and quality is still challenging. Especially, AI/GenAI resource-constrained enterprises must deal with complex domainspecific technical documents of assets and designs supplied by multiple vendors. This paper presents novel practical methods that incorporate contexts into the design and execution of activities for industrial technical document extraction applications. We consider resource-constrained environments in which enterprises are with edge GenAI/LLMs infrastructures and non-AI engineers. We devise context-aware composition and adaptation for extraction pipelines to deal with diverse attributes of GenAI/LLMs and extraction quality control. We experiment our methods with technical documents for telco operators.

Swarm-edge-cloud service-based applications (SESA) utilize UAV swarm nodes and edge-cloud computing infrastructures to perform complex tasks. These applications leverage distributed services across UAVs and edge resources, transitioning to cloud resources as needed, to support continuum computing, adapting to dynamic workloads, network conditions, and mission requirements. However, existing solutions lack robust mechanisms for dynamic policy enforcement in such environments. This paper introduces SAGELY (SwArm-edGE PoLicY), a framework for secure, efficient, and adaptive policy enforcement in dynamic SESA environments. SAGELY incorporates: (1) context-aware policy adaptation to adjust enforcement dynamically, (2) flexible policy enforcement across centralized and decentralized models, and (3) a pluggable service architecture for continuum context management. We present a testbed to enforce continuum policies and conduct experiments.

The Internet of Things (IoT) is becoming increasingly ubiquitous, acting as an important source of real-time data for various applications. By allowing data exchange between various parties along the IoT devices-Edge-Cloud computing continuum, the larger societal benefits of the IoT can be achieved. Assuring security and fostering confidence for IoT data sharing, however, is one of the biggest obstacles. Sharing real-time data originating from connected devices is crucial to real-world intelligent IoT applications, i.e., based on artificial intelligence/machine learning. Such IoT data sharing involves multiple parties for different purposes and is usually based on data contracts that might depend on the dynamic change of IoT data variety and velocity. We aim to support multiple parties (aka tenants) with dynamic contracts based on the data value for their specific contextual purposes. This work addresses these challenges by introducing a novel dynamic context-based policy enforcement framework to support IoT data sharing (on-Edge) based on dynamic contracts. Our enforcement framework allows IoT Data Hub owners to define extensible rules and metrics to govern the tenants accessing the shared data on the Edge based on policies defined with static and dynamic contexts. We have created an edge-centered architecture that enables multi-tenant use cases with tenant-specific application deployment and IoT-context-based data sharing on edge servers. Our proof-of-concept prototype for sharing sensitive data such as surveillance camera videos has illustrated our proposed framework. The experimental results demonstrated that our framework could soundly and timely enforce context-based policies at runtime with moderate overhead. Moreover, the context and policy changes are correctly reflected in the system in nearly real-time. We have addressed the need to enable multi-parties IoT (data) resources to be shared based on contracts, especially with dynamic IoT contexts, for tenant applications on the edge to allow their closer access to data.

Deploying end-to-end ML applications on edge resources becomes a viable solution to achieve performance and data regulations. With the microservice architecture, these applications can scale dynamically, improving service availability under dynamic workloads. However, orchestrating multiple end-to-end ML applications within heterogeneous edge environments must deal with numerous challenges while sharing computing resources. Prevalent orchestration tools/frameworks supporting edge ML serving are inefficient in provisioning methods due to constrained resources, diverse resource demands and utilization patterns. In this work, we present a provisioning method to optimize resource utilization for end-to-end ML applications on a heterogeneous edge. By profiling all microservices within the application, we estimate scales and allocate them on desired hardware platforms with sufficient resources when considering their runtime utilization patterns. We also provide several practical analyses on runtime monitoring metrics to detect and mitigate resource contentions, guaranteeing performance. The experiments with three real-world ML applications demonstrate the practicality of our method on a heterogeneous edge cluster of Raspberry Pis and Jetson Developer Kits.

Recent advances in AI/ML technologies have accelerated the development of various ML applications. One of the major trends in AI/ML application development is the increasing use of multiple ML models to support high-accuracy inference in a complex end-to-end ML serving. However, testing the right configuration of multiple ML models is expensive, and the application requirements for ML inferences are highly dependent on various factors like the quality of ML models, computing resource performance, and data quality. In this context, techniques and methods that help to emulate and analyze ML inference characteristics using queueing theory can reduce the development effort and cost for ML services encapsulating ML models but also the entire ML system. In this paper, we modeled and analyzed a queueing model for an ML system that uses ensemble learning as an inference method with a new rule and clarified the impacts of model design in ensemble learning on the system's performance. As a result, we demonstrate the usefulness of the analysis for understanding possible configurations and their efficiency in the ML system through queueing analysis and simulation.

A long running data-intensive computational application acquires costly computing resources. With the emerging new architectures, like computing systems with multiple nodes of many-core CPUs and accelerators, while domain-specific tools and libraries employed in such an application leverage high parallelism on accelerators for intensive computations, the remaining resources can potentially be utilized for other application-related data operations. Such data operations, called opportunistic data operations in this work, must usually be carried out for post-processing or follow-up analytics based on results produced during the runtime of the application. These operations are not easily backfilled or preempted under the guidance of the domain scientist or by common task scheduling systems due to their complex dependencies.In this paper, we introduce a framework for domain scientists to identify and execute opportunistic data operation tasks. With a minimal specification or modification of the main application, the scientists can specify, monitor, and execute opportunistic tasks independently from the main application and the framework will detect underutilized resources to execute these tasks, thereby, optimizing utilization efficiency within the allocated resources. We present experiments to demonstrate the applicability of our framework on a magnetic field modeling running on the LUMI computing system.

The increasing development and utilization of Large Language Model (LLM) services have demonstrated many benefits in different contexts. However, LLM services are mainly available in the public cloud and require huge computing resources to operate, thus not accessible to many companies, organizations or communities with constrained resources. While research efforts have concentrated on LLMs quantization for resource-constrained computing environments like edge devices, to democratize the availability of LLM services as utilities for such communities requires much more than the optimization of LLM models. In this paper, we introduce CULAO - a framework for constructing utilities from LLMs in resource-constrained environments. Our framework focuses on key requirements of resource-constrained companies, organizations and communities by enabling the provisioning and coordination of LLMs as utilities, based on the availability of open-source LLMs.

Optimizing the quality of machine learning (ML) services for individual consumers with specific objectives is crucial for improving consumer satisfaction. In this context, end-to-end ensemble ML serving (EEMLS) faces many challenges in selecting and deploying ensembles of ML models on diverse resources across the edge-cloud continuum. This paper provides a method for evaluating the runtime performance of inference services via consumer-defined metrics. We enable ML consumers to define high-level metrics and consider consumer satisfaction in estimating service costs. Moreover, we introduce a time-efficient ensemble selection algorithm to optimize the EEMLS with intricate trade-offs between service quality and costs. Our intensive experiments demonstrate that the algorithm can be executed periodically despite the extensive search space, enabling dedicated optimization for individual consumers in dynamic contexts.

Emerging generative and fine-tuning LLMs services have been widely benchmarked and used for various software development tasks. These LLMs services are powerful but have different output qualities for software development tasks and may not be able to deal with complex development tasks in edge-cloud software modernization and new developments due to their generative capabilities and lack of up-to-date (domain) knowledge. Many queries and solutions related to target platforms, deployment configurations, policies, data regulation, observability, to name just a few, are not well integrated with these LLMs, but are accessed by the developer through other sources. In this work, we discuss situations where the gaps between the needs and the offerings from LLMs can be compensated by Platform Knowledge, which captures knowledge about, e.g., software, service and infrastructure catalogs, architectural decision records and code patterns. We propose COLLMS - a framework for coordinating LLMs services and Platform Knowledge. At the starting point of the framework, we will discuss challenges for achieving the coordination centered around Platform Knowledge, LLMs management and integration, quality-aware coordination of LLMs, and observability and knowledge updating.

Given a large-scale mobile network with a variety of equipment and radio access network technologies for an approximate 20 million subscribers, there are many types of data that can be used for big data analytics and machine learning (ML) tasks for network operations, monitoring, and optimization. However, a variety of data is measured, collected, and propagated through numerous complex data and software systems. Thus, people, software components, and data-driven operations for big data and ML pipelines face great challenges in dealing with data quality impacts. Data quality related problems occur and are propagated through complex operations involving different types of data, people, software components, and analytics that cannot be solved purely through data quality engineering. This article discusses our TENSAI framework, as a practical and responsible observability for ensuring data quality in such a mobile network. TENSAI focuses on methods of communication, strategy specifications, and data quality engineering for diverse types of data and analytics among different types of operations. TENSAI presents techniques for capturing and communicating causes/effects of data quality problems clearly to all relevant stakeholders, developing data quality-aware adaptation strategies for actions on data that can be integrated into analytics processes, and engineering data quality awareness in software and data pipelines. Thus, TENSAI supports full visibility of data quality problems and impacts among related systems to empower the utilization and adaptation of data analytics for different types of operations. We will illustrate our TENSAI with several real-world data types, pipelines, and cases based on a real-world mobile network.

The resilience of process-aware information systems (PAIS) is vital for enterprises' competitiveness in the ever-changing world. Enterprises need continuously maintain business processes in PAIS against uncertainty, resulting in the constant research topic of change management in business process management (BPM). However, previous efforts have left some gaps in dealing with uncertainties in change requirements, either due to high costs or technical incapability. Among these gaps, long-tailed change (LTC), characterized as the urgent and customizable maintenance needs in response to the residual uncertainty, has been largely undeveloped. COVID-19 and some emergent events in 2022 call for more effective solutions to such problems, to mitigate the business loss and to seek emerging opportunities. LTC challenges come from the contention between the degree of flexibility and compliance with business regulations in the context of rapid adaptation. In this paper, we develop a systematic approach to deal with these issues. First, we formulate the problem as a cooperation of multiple participants in the business by leveraging the separation of concerns principle to clarify the division of labor contributing to agility and fine-tuning capability of the approach, and to provide a controllable mechanism to balance operational resilience and dependable adaptation for PAIS. According to the framework, we develop a domain-specific language (DSL) and corresponding techniques to support business people in customizing business processes' behavior to meet circumstances. At the same time, business stakeholders can enforce core functions and service-level agreement (SLA) constraints on the whole business process. We validate our framework through two scenarios, one for design-time model adaptation and the other for run-time instance adaptation. These experiments reveal promising potentials of the framework in adapting to LTC and resolving the contention between agility/flexibility in change and compliance with business regulations, which significantly improves PAIS's resilience to wide-spectrum, low-frequent, sporadic, and transitory events.

This paper presents our approach for supporting machine learning (ML)-based analytics of quality of experience (QoE) related issues in a variety of Radio Access Networks (V-RAN). We focus on key problems in a holistic analytics infrastructure for engineers without strong ML skills and powerful computing infrastructures. We characterize types of relevant data and existing data systems to follow a specific data mesh approach suitable for engineers. The paper presents key steps in establishing the participation of engineers and the acquisition of domain knowledge. We introduce models for representing analytics subjects and their dependencies, and for managing relevant ML techniques and methods for analytics subjects. We explain our work through examples from a large-scale mobile network of approximately 4 million subscribers.

The maturity of machine learning (ML) development and the decreasing deployment cost of capable edge devices have proliferated the development and deployment of edge ML solutions for critical IoT-based business applications. The combination of edge computing and ML not only addresses the development cost barrier, but also solves the obstacles due to the lack of powerful cloud data centers. However, not only the edge ML research and development is still at an early stage and requires substantial skills normally missed in resource-constrained communities, but also various infrastructure constraints w.r.t. network reliability and computing power, and business contexts from the resource-constrained environments require different considerations to make edge ML applications context aware through smart and intelligent runtime strategies. In this paper, we analyze representative real-world business scenarios for edge ML solutions and their contexts in resource-constrained communities and environments. We identify and map the key distinguished contexts of distributed edge ML and discuss the impacts of these contexts on data and software components and deployment models. Finally, we present key research areas, how we should approach them, and possible tooling for making edge machine learning solutions smarter in resource-constrained communities and environments.

Sharing real-time data originating from connected devices is crucial to real-world intelligent Internet of Things (IoT) applications, i.e., based on artificial intelligence/machine learning (AI/ML). Such IoT data sharing involves multiple parties for different purposes and is usually based on data contracts that might depend on the dynamic change of IoT data variety and velocity. It is still an open challenge to support multiple parties (aka tenants) with these dynamic contracts based on the data value for their specific contextual purposes.This work addresses these challenges by introducing a novel dynamic context-based policy enforcement framework to support IoT data sharing (on-Edge) based on dynamic contracts. Our enforcement framework allows IoT Data Hub owners to define extensible rules and metrics to govern the tenants in accessing the shared data on the Edge based on policies defined with static and dynamic contexts. We have developed a proof-of-concept prototype for sharing sensitive data such as surveillance camera videos to illustrate our proposed framework. The experimental results demonstrated that our framework could soundly and timely enforce context-based policies at runtime with moderate overhead. Moreover, the context and policy changes are correctly reflected in the system in nearly real-time.

The origin of the quiet Sun magnetism is under debate. Investigating the solar cycle variation observationally in more detail can give us clues about how to resolve the controversies. We investigate the solar cycle variation of the most magnetically quiet regions and their surface gravity oscillation ($f$-) mode integrated energy ($E_f$). We use 12 years of HMI data and apply a stringent selection criteria, based on spatial and temporal quietness, to avoid any influence of active regions (ARs). We develop an automated high-throughput pipeline to go through all available magnetogram data and to compute $E_f$ for the selected quiet regions. We observe a clear solar cycle dependence of the magnetic field strength in the most quiet regions containing several supergranular cells. For patch sizes smaller than a supergranular cell, no significant cycle dependence is detected. The $E_f$ at the supergranular scale is not constant over time. During the late ascending phase of Cycle 24 (SC24, 2011-2012), it is roughly constant, but starts diminishing in 2013, as the maximum of SC24 is approached. This trend continues until mid-2017, when hints of strengthening at higher southern latitudes are seen. Slow strengthening continues, stronger at higher latitudes than at the equatorial regions, but $E_f$ never returns back to the values seen in 2011-2012. Also, the strengthening trend continues past the solar minimum, to the years when SC25 is already clearly ascending. Hence the $E_f$ behavior is not in phase with the solar cycle. The anticorrelation of $E_f$ with the solar cycle in gross terms is expected, but the phase shift of several years indicates a connection to the poloidal large-scale magnetic field component rather than the toroidal one. Calibrating AR signals with the QS $E_f$ does not reveal significant enhancement of the $f$-mode prior to AR emergence.

As blockchain becomes an essential part of many software systems in the edge and cloud, the developer starts to treat blockchain features like commodity software components that can be integrated into edge and cloud software systems. For the developer it is quite challenging to determine, customize, and evaluate suitable blockchain features for software systems in the edge and cloud environments. In this paper, we conceptualize important blockchain interactions in mobile edge computing software systems (MECSS) and present generic techniques for evaluating these interactions. We determine different interaction patterns for different deployments of compute resources and networks. We abstract and represent application-level mobile edge computing (MEC) features and blockchain features to create MECSS deployment models to be coupled with testbed deployments for benchmarking application-level interactions within application contexts. Based on that, we develop a generic framework for building and executing benchmarks of applicationlevel blockchain interactions within MECSS. We will demonstrate our framework for vehicle-to-everything communication scenarios with two main blockchain technologies, Hyperledger Fabric and Ethereum, using various types of compute resources in edge and cloud infrastructures.