Searcharxiv⌕ Search

arXiv subjects

Marin Litoiu

Publications and source records attributed to Marin Litoiu.

15 recordsLinked to original sources

Adaptive Multi-Discriminator WGAN Framework for Resource-Constrained Internet of Vehicles Using Reinforcement Learning and Game Theory

Managing machine learning workloads as a network service introduces a resource-orchestration problem distinct from conventional model training; which nodes should be allocated to a task, how communication and computation budgets should be divided among them, and how service quality should be sustained as connectivity and node availability change with mobility. Deploying Generative Adversarial Networks (GANs) in Internet of Vehicles (IoV) environments is a demanding instance of this problem; resource constraints, dynamic network topologies, and competing optimization objectives mean that traditional GAN architectures cannot simultaneously achieve high accuracy, efficient resource use, low delay, and low communication overhead. This paper introduces an adaptive multi-discriminator Wasserstein GAN (MD-WGAN) framework that integrates reinforcement learning with game-theoretic coordination to address these challenges jointly. In our framework, roadside units host generators paired with Deep Q-Network (DQN) agents that select discriminator subsets and manage distributed training across mobile vehicular nodes, while a game-theoretic coordination step allocates training epochs between generators and discriminators. A unified optimization objective ties adversarial learning quality to resource efficiency, communication overhead, and latency under vehicular constraints, allowing the framework to continuously adapt its training behavior as network conditions change. Evaluation on real-world NGSIM trajectory data shows that the framework attains prediction accuracy comparable to state-of-the-art GAN baselines - the lowest RMSE (1.029) and MAE (0.894) among all evaluated methods - while markedly improving resource efficiency: average CPU utilization is reduced by roughly 28% and mean memory usage by roughly 6%, at competitive communication overhead and latency.

cs.LG↗

Mitigating Uncertainty Interactions in GenAI-based Adaptive Systems: Vision, Challenges and Preliminary Guidelines

Modern software-intensive systems increasingly incorporate GenAI components, including LLM and agentic subsystems, which introduce novel and compounding sources of uncertainty across system layers. The self-adaptive systems community has made significant strides in understanding and managing uncertainty. The intrinsic characteristics of GenAI, including probabilistic outputs, hallucinations, context window limitations, and memory staleness, demand a re-examination of existing frameworks and mitigation strategies for dealing with uncertainty, and especially, the interactions among uncertainty occurrences and manifestations. This position paper posits an initial conceptual framework that establishes preliminary guidelines for characterizing and mitigating uncertainty interactions in GenAI-based software-intensive systems. We argue that mitigation must be considered across the full software lifecycle, encompassing requirements analysis, design-time decisions, and run-time adaptations, which implies the unfeasibility of decoupling it from monitoring and analysis at execution. We introduce the notion of a mitigation function that maps uncertainty interaction categories, involved uncertainty types, system design characteristics, criticality, and mitigating architectural tactics to impacted quality attributes and corresponding tradeoffs. As a proof of concept, we apply the framework to three illustrative use cases: LLM serving, AI-driven software operations, and agentic cloud management.

cs.SE↗

SFlexRCA: Lightweight, Scalable, and Flexible Root Cause Analysis for IIoT Edge Systems

Industrial Internet of Things (IIoT) systems generate high-dimensional sensor telemetry from interconnected components, where faults can propagate across the system. To address these challenges, we propose SFlexRCA (Scalable and Flexible Root Cause Analysis), a topology-free RCA framework designed for resource-constrained IIoT environments. SFlexRCA transforms multivariate telemetry into compact orthogonal representations and applies shared lightweight linear modeling, avoiding explicit graph construction, message passing, and per-variable or lag-specific parameter growth. We evaluate SFlexRCA on three publicly available IIoT datasets, BATADAL, SWaT, and WADI, spanning different numbers of monitored variables, temporal characteristics, and training-data regimes. SFlexRCA is compared with 10 statistical, causal, and non-causal baselines in terms of RCA accuracy, training efficiency, inference latency, and memory consumption. In addition, inference efficiency and memory consumption are evaluated on Raspberry Pi 3 and Raspberry Pi 5, while energy consumption is additionally measured on Raspberry Pi 5. We further investigate temporalcontext sensitivity, architectural and loss components, and alternative representations. Notably, SFlexRCA maintains strong localization performance on BATADAL despite its limited normal-operation training data, while its compact shared architecture avoids the parameter growth associated with causal and graph-based approaches. Its lightweight shared architecture further enables efficient deployment on resource-constrained IIoT edge devices. The SFlexRCA code is available at https://github.com/theamrzaki/RootCause- Analysis-Correlation-Attentive-Modeling.

cs.LG↗

OrEdge: Efficient Multi-Modal Anomaly Detection in Distributed Software Systems via Orthogonal-Domain Learning

We introduce Orthogonal-Edge (OrEdge), a lightweight framework for real-time anomaly detection in multi-modal distributed software systems. Unlike existing approaches that rely on computationally expensive attention- and graph-based architectures, OrEdge leverages orthogonal-domain temporal representations to achieve accurate anomaly detection with substantially lower computational complexity and model size. It jointly analyzes heterogeneous monitoring data, including logs, metrics, and traces, to identify abnormal software behavior, capture temporal dependencies, and reduce redundancy across observability signals. At its core, OrEdge incorporates OrEdgeCore, a lightweight orthogonal-domain reconstruction module that captures recurring temporal patterns while suppressing transient variations. Evaluated on three real-world microservice datasets (MSDS, SN, and TT), OrEdge achieves competitive detection performance while reducing the reconstruction model size to at most 9.6K parameters, compared with 20K--143K parameters in existing methods. This compact design enables efficient deployment on resource-constrained edge devices: on Raspberry Pi platforms, OrEdge achieves sub-second inference and reduces inference latency by over an order of magnitude compared with existing approaches. Extensive ablation studies, sensitivity analyses, orthogonal basis evaluations, and qualitative case studies further validate the effectiveness of each design component. Overall, OrEdge demonstrates that orthogonal-domain temporal modeling provides an effective alternative to computationally intensive attention- and graph-based architectures, achieving a favorable balance between detection accuracy and computational efficiency for real-time multi-modal anomaly detection in edge environments. The code is available at https://github.com/theamrzaki/MicroService_Twin_Original.

cs.SE↗

Towards Sustainable Large Language Model Inference: A Systematic Literature Review

Large language model (LLM) inference is becoming a significant contributor to both operational cost and environmental impact in modern AI deployments. Energy consumption during inference depends on multiple factors across the pipeline, including input and output length, model design, system-level decisions, and hardware characteristics. This paper presents a systematic literature review (SLR) of energy-efficient and sustainable LLM inference. We analyze selected primary studies to identify techniques proposed to reduce energy consumption across different levels of the inference pipeline, including input, model, system, and hardware. We further examine the trade-offs between energy consumption, latency, and accuracy, highlighting inconsistencies in how these relationships are evaluated. In addition, we review the tools, datasets, and metrics used for energy measurement, covering both software-based and hardware-based approaches. This survey provides a comprehensive synthesis of existing literature, offering insights into the effectiveness of different techniques in terms of energy savings and their associated impact on accuracy. It highlights the trade-offs between energy efficiency and performance and identifies limitations in current evaluation practices. Overall, this work aims to support future research toward more efficient and environmentally responsible deployment of LLMs. Our repository of relevant references is maintained on GitHub.

cs.SE↗

Employing Software Diversity in Cloud Microservices to Engineer Reliable and Performant Systems

In the ever-shifting landscape of software engineering, we recognize the need for adaptation and evolution to maintain system dependability. As each software iteration potentially introduces new challenges, from unforeseen bugs to performance anomalies, it becomes paramount to understand and address these intricacies to ensure robust system operations during the lifetime. This work proposes employing software diversity to enhance system reliability and performance simultaneously. A cornerstone of our work is the derivation of a reliability metric. This metric encapsulates the reliability and performance of each software version under adverse conditions. Using the calculated reliability score, we implemented a dynamic controller responsible for adjusting the population of each software version. The goal is to maintain a higher replica count for more reliable versions while preserving the diversity of versions as much as possible. This balance is crucial for ensuring not only the reliability but also the performance of the system against a spectrum of potential failures. In addition, we designed and implemented a diversity-aware autoscaling algorithm that maintains the reliability and performance of the system at the same time and at any scale. Our extensive experiments on realistic cloud microservice-based applications show the effectiveness of the proposed approach in this paper in promoting both reliability and performance.

cs.SE↗

A Learning-Based Caching Mechanism for Edge Content Delivery

With the advent of 5G networks and the rise of the Internet of Things (IoT), Content Delivery Networks (CDNs) are increasingly extending into the network edge. This shift introduces unique challenges, particularly due to the limited cache storage and the diverse request patterns at the edge. These edge environments can host traffic classes characterized by varied object-size distributions and object-access patterns. Such complexity makes it difficult for traditional caching strategies, which often rely on metrics like request frequency or time intervals, to be effective. Despite these complexities, the optimization of edge caching is crucial. Improved byte hit rates at the edge not only alleviate the load on the network backbone but also minimize operational costs and expedite content delivery to end-users. In this paper, we introduce HR-Cache, a comprehensive learning-based caching framework grounded in the principles of Hazard Rate (HR) ordering, a rule originally formulated to compute an upper bound on cache performance. HR-Cache leverages this rule to guide future object eviction decisions. It employs a lightweight machine learning model to learn from caching decisions made based on HR ordering, subsequently predicting the "cache-friendliness" of incoming requests. Objects deemed "cache-averse" are placed into cache as priority candidates for eviction. Through extensive experimentation, we demonstrate that HR-Cache not only consistently enhances byte hit rates compared to existing state-of-the-art methods but also achieves this with minimal prediction overhead. Our experimental results, using three real-world traces and one synthetic trace, indicate that HR-Cache consistently achieves 2.2-14.6% greater WAN traffic savings than LRU. It outperforms not only heuristic caching strategies but also the state-of-the-art learning-based algorithm.

cs.NI↗

LearnedWMP: Workload Memory Prediction Using Distribution of Query Templates

In a modern DBMS, working memory is frequently the limiting factor when processing in-memory analytic query operations such as joins, sorting, and aggregation. Existing resource estimation approaches for a DBMS estimate the resource consumption of a query by computing an estimate of each individual database operator in the query execution plan. Such an approach is slow and error-prone as it relies upon simplifying assumptions, such as uniformity and independence of the underlying data. Additionally, the existing approach focuses on individual queries separately and does not factor in other queries in the workload that may be executed concurrently. In this research, we are interested in query performance optimization under concurrent execution of a batch of queries (a workload). Specifically, we focus on predicting the memory demand for a workload rather than providing separate estimates for each query within it. We introduce the problem of workload memory prediction and formalize it as a distribution regression problem. We propose Learned Workload Memory Prediction (LearnedWMP) to improve and simplify estimating the working memory demands of workloads. Through a comprehensive experimental evaluation, we show that LearnedWMP reduces the memory estimation error of the state-of-the-practice method by up to 47.6%. Compared to an alternative single-query model, during training and inferencing, the LearnedWMP model and its variants were 3x to 10x faster. Moreover, LearnedWMP-based models were at least 50% smaller in most cases. Overall, the results demonstrate the advantages of the LearnedWMP approach and its potential for a broader impact on query performance optimization.

cs.DB↗

Self-Adaptation in Industry: A Survey

Computing systems form the backbone of many areas in our society, from manufacturing to traffic control, healthcare, and financial systems. When software plays a vital role in the design, construction, and operation, these systems are referred as software-intensive systems. Self-adaptation equips a software-intensive system with a feedback loop that either automates tasks that otherwise need to be performed by human operators or deals with uncertain conditions. Such feedback loops have found their way to a variety of practical applications; typical examples are an elastic cloud to adapt computing resources and automated server management to respond quickly to business needs. To gain insight into the motivations for applying self-adaptation in practice, the problems solved using self-adaptation and how these problems are solved, and the difficulties and risks that industry faces in adopting self-adaptation, we performed a large-scale survey. We received 184 valid responses from practitioners spread over 21 countries. Based on the analysis of the survey data, we provide an empirically grounded overview of state-of-the-practice in the application of self-adaptation. From that, we derive insights for researchers to check their current research with industrial needs, and for practitioners to compare their current practice in applying self-adaptation. These insights also provide opportunities for the application of self-adaptation in practice and pave the way for future industry-research collaborations.

cs.SE↗

Building Automation System Data Integration with BIM: Data Structure and Supporting Case Study

Buildings Automation Systems (BAS) are ubiquitous in contemporary buildings, both monitoring building conditions and managing the building system control points. At present, these controls are prescriptive and pre-determined by the design team, rather than responsive to actual building performance. These are further limited by prescribed logic, possess only rudimentary visualizations, and lack broader system integration capabilities. Advances in machine learning, edge analytics, data management systems, and Facility Management-enabled Building Information Models (FM-BIMs) permit a novel approach: cloud-hosted building management. This paper presents an integration technique for mapping the data from a building Internet of Things (IoT) sensor network to an FM-BIM. The sensor data naming convention and timeseries analysis strategies integrated into the data structure are discussed and presented, including the use of a 3D nested list to permit timeseries data to be mapped to the FM-BIM and readily visualized. The developed approach is presented through a case study of an office living lab consisting of a local sensor network mimicking a BAS, which streams to a cloud server via a virtual private network connection. The resultant data structure and key visualizations are presented to demonstrate the value of this approach, which permits the end-user to select the desired timeframe for visualization and readily step through the spatio-temporal building performance data.

cs.NI↗

Has Your FaaS Application Been Decommissioned Yet? -- A Case Study on the Idle Timeout in Function as a Service Infrastructure

Function as a Service (FaaS) is a new cloud technology with automated resource management. Different from traditional cloud computing, each FaaS cloud function can only run a fixed period of time before being decommissioned. Furthermore, FaaS cloud providers often update their platforms (e.g., idle timeout). These changes and their associated impact are not transparent and could potentially impact the execution of the cloud functions. Hence, in this paper, we develop a methodology to characterize the cloud function idle timeout which is the duration a FaaS cloud provider keeps a cloud function instance alive without serving active traffic. Our study was conducted on three popular FaaS platforms, namely AWS Lambda, IBM and Azure Cloud Function. Moreover, we also report how long a cloud function instance can be kept alive when a user regularly polls the instance. Experimental results show that the idle timeout period has evolved from 01/2020 till 01/2022.

cs.DC↗

Towards Better Adaptive Systems by Combining MAPE, Control Theory, and Machine Learning

Two established approaches to engineer adaptive systems are architecture-based adaptation that uses a Monitor-Analysis-Planning-Executing (MAPE) loop that reasons over architectural models (aka Knowledge) to make adaptation decisions, and control-based adaptation that relies on principles of control theory (CT) to realize adaptation. Recently, we also observe a rapidly growing interest in applying machine learning (ML) to support different adaptation mechanisms. While MAPE and CT have particular characteristics and strengths to be applied independently, in this paper, we are concerned with the question of how these approaches are related with one another and whether combining them and supporting them with ML can produce better adaptive systems. We motivate the combined use of different adaptation approaches using a scenario of a cloud-based enterprise system and illustrate the analysis when combining the different approaches. To conclude, we offer a set of open questions for further research in this interesting area.

cs.SE↗

Performance Modeling of Microservice Platforms

Microservice architecture has transformed the way developers are building and deploying applications in the nowadays cloud computing centers. This new approach provides increased scalability, flexibility, manageability, and performance while reducing the complexity of the whole software development life cycle. The increase in cloud resource utilization also benefits microservice providers. Various microservice platforms have emerged to facilitate the DevOps of containerized services by enabling continuous integration and delivery. Microservice platforms deploy application containers on virtual or physical machines provided by public/private cloud infrastructures in a seamless manner. In this paper, we study and evaluate the provisioning performance of microservice platforms by incorporating the details of all layers (i.e., both micro and macro layers) in the modelling process. To this end, we first build a microservice platform on top of Amazon EC2 cloud and then leverage it to develop a comprehensive performance model to perform what-if analysis and capacity planning for microservice platforms at scale. In other words, the proposed performance model provides a systematic approach to measure the elasticity of the microservice platform by analyzing the provisioning performance at both the microservice platform and the back-end macroservice infrastructures.

cs.DC↗

Understanding Brain Dynamics for Color Perception using Wearable EEG headband

The perception of color is an important cognitive feature of the human brain. The variety of colors that impinge upon the human eye can trigger changes in brain activity which can be captured using electroencephalography (EEG). In this work, we have designed a multiclass classification model to detect the primary colors from the features of raw EEG signals. In contrast to previous research, our method employs spectral power features, statistical features as well as correlation features from the signal band power obtained from continuous Morlet wavelet transform instead of raw EEG, for the classification task. We have applied dimensionality reduction techniques such as Forward Feature Selection and Stacked Autoencoders to reduce the dimension of data eventually increasing the model's efficiency. Our proposed methodology using Forward Selection and Random Forest Classifier gave the best overall accuracy of 80.6\% for intra-subject classification. Our approach shows promise in developing techniques for cognitive tasks using color cues such as controlling Internet of Thing (IoT) devices by looking at primary colors for individuals with restricted motor abilities.

cs.LG↗

Using Models at Runtime to Address Assurance for Self-Adaptive Systems

A self-adaptive software system modifies its behavior at runtime in response to changes within the system or in its execution environment. The fulfillment of the system requirements needs to be guaranteed even in the presence of adverse conditions and adaptations. Thus, a key challenge for self-adaptive software systems is assurance. Traditionally, confidence in the correctness of a system is gained through a variety of activities and processes performed at development time, such as design analysis and testing. In the presence of selfadaptation, however, some of the assurance tasks may need to be performed at runtime. This need calls for the development of techniques that enable continuous assurance throughout the software life cycle. Fundamental to the development of runtime assurance techniques is research into the use of models at runtime (M@RT). This chapter explores the state of the art for usingM@RT to address the assurance of self-adaptive software systems. It defines what information can be captured by M@RT, specifically for the purpose of assurance, and puts this definition into the context of existing work. We then outline key research challenges for assurance at runtime and characterize assurance methods. The chapter concludes with an exploration of selected application areas where M@RT could provide significant benefits beyond existing assurance techniques for adaptive systems.

cs.SE↗