Distributed Service Architecture Research Articles

Research on the Construction of an Efficient and Lightweight Online Detection Method for Tiny Surface Defects through Model Compression and Knowledge Distillation

In response to the current issues of poor real-time performance, high computational costs, and excessive memory usage of object detection algorithms based on deep convolutional neural networks in embedded devices, a method for improving deep convolutional neural networks based on model compression and knowledge distillation is proposed. Firstly, data augmentation is employed in the preprocessing stage to increase the diversity of training samples, thereby improving the model’s robustness and generalization capability. The K-means++ clustering algorithm generates candidate bounding boxes, adapting to defects of different sizes and selecting finer features earlier. Secondly, the cross stage partial (CSP) Darknet53 network and spatial pyramid pooling (SPP) module extract features from the input raw images, enhancing the accuracy of defect location detection and recognition in YOLO. Finally, the concept of model compression is integrated, utilizing scaling factors in the batch normalization (BN) layer, and introducing sparse factors to perform sparse training on the network. Channel pruning and layer pruning are applied to the sparse model, and post-processing methods using knowledge distillation are used to effectively reduce the model size and forward inference time while maintaining model accuracy. The improved model size decreases from 244 M to 4.19 M, the detection speed increases from 32.8 f/s to 68 f/s, and mAP reaches 97.41. Experimental results demonstrate that this method is conducive to deploying network models on embedded devices with limited GPU computing and storage resources. It can be applied in distributed service architectures for edge computing, providing new technological references for deploying deep learning models in the industrial sector.

Adaptive exchange protocol for multi-agent communication in augmented reality system

Augmented Reality (AR) is one of the modern ways of providing users with different types of information. In applications based on this technology, the main influence on the subjective assessment of product quality is the speed of interaction between the system and the user and between users of the system. The main purpose of the work series are creation and modification for algorithms to assessing and improving the quality of AR services. This paper describes and justifies an adaptive communication protocol for multi-agent interaction. We consider a general nonlinear dynamic system with introducing feedback control which is based on measurements under almost arbitrary noise. In practice, this control is realized as a superposition of neural networks with the estimation of the result of mutual additional learning based on the system identification method (M.K.Campi and E.Weyer’s LSCR is used). The prototype is built for a system with a distributed server architecture and multiagent behavior of both clients and servers in the system.

Adaptively secure multi-authority attribute-based broadcast encryption in fog computing

Fog computing is considered an important development in cloud computing, with shorter response times, less data transfer load and a more secure distributed service architecture that can be accessed by a wide range of end devices. As with cloud computing, it also faces the challenge of how to securely distribute sensitive data. Ciphertext-Policy Attribute-based Broadcast Encryption (CP-ABBE) can protect sensitive data and provide fine-grained access control with user revocation. However, existing CP-ABBE schemes suffer many limitations when implemented in a fog computing environment: a single authority performs poorly in handling requests or data from many IoT devices, leading to system delays or limited functionality; expensive decryption costs make the computation and storage overheads unfriendly to devices with limited resources; these schemes are only proven selectively secure, meaning that security was proven in a weaker model. In this paper, using threshold secret sharing and decryption delegation, we propose an adaptively secure attribute-based multi-authority broadcast encryption (MA-ABBE) scheme that breaks through these limitations. The comparison and analysis results show that our scheme is practical and achieves adaptive security, enables user-side fast decryption and low storage overhead, and provides a secure and efficient option for protecting sensitive data in fog computing environments.

BGSD: A SBERT and GAT-based Service Discovery Framework for Heterogeneous Distributed IoT

Collaboration between heterogeneous IoT (Internet of Things) devices can enrich the capabilities of the system. In scenarios where network infrastructure or cloud platforms are unavailable, enabling devices to collaborate autonomously in a distributed manner can improve the scalability and invulnerability of the system. A distributed service architecture with semantic services can enable more flexible collaboration among IoT devices, while the efficiency and accuracy of current methods on devices with poor computing and storage capacity need to be improved. Therefore, this paper proposes a SBERT (Sentence Bidirectional Encoder Representations from Transformers) and GAT (Graph Attention Networks) based Service Discovery method, called BGSD. SBERT is used to extract deep semantic features of services during offline to improve the discriminability of semantic features and reduce performance requirements for IoT devices, and GAT is used to perceive the local topology environment to improve the network navigability of service search. To evaluate the method, we generate a heterogeneous IoT topology dataset based on OWLS-TC4. The simulation results of service discovery show that BGSD has advantages in search efficiency, matching accuracy and computational cost, and can support the collaboration of IoT devices in heterogeneous distributed scenarios, which verifies the rationality of this method.

Digital Transformation in Metrology: Building a Metrological Service Ecosystem

In this paper a metrological quality infrastructure is introduced, that considers the strict legal framework and offers transparency, security and resilience. As a foundation, the AnGeWaNt platform is used to connect external infrastructures and foster digital transformation of processes. A fast and easy integration of a vacuum laboratory workflow is demonstrated by benefiting from the distributed service architecture and harmonized RESTful application interfaces. Furthermore, through the modular approach of the AnGeWaNt platform, a rapid service prototyping is enabled by a tool set of services. Providing a digital transformed conformity process to the service hub and avoiding expensive re-implementation. These two exemplary use cases are extensively described and their benefits to the metrological service ecosystem are shown. The main advantage of the interdependent service ecosystem is the flexibility to preserve specialized applications on site, while enabling innovative and custom-tailored services at the same time.

DCC Terminology Service—An Automated CI/CD Pipeline for Converting Clinical and Biomedical Terminologies in Graph Format for the Swiss Personalized Health Network

One goal of the Swiss Personalized Health Network (SPHN) is to provide an infrastructure for FAIR (Findable, Accessible, Interoperable and Reusable) health-related data for research purposes. Semantic web technology and biomedical terminologies are key to achieving semantic interoperability. To enable the integrative use of different terminologies, a terminology service is a important component of the SPHN Infrastructure for FAIR data. It provides both the current and historical versions of the terminologies in an SPHN-compliant graph format. To minimize the usually high maintenance effort of a terminology service, we developed an automated CI/CD pipeline for converting clinical and biomedical terminologies in an SPHN-compatible way. Hospitals, research infrastructure providers, as well as any other data providers, can download a terminology bundle (currently composed of SNOMED CT, LOINC, UCUM, ATC, ICD-10-GM, and CHOP) and deploy it in their local terminology service. The distributed service architecture allows each party to fulfill their local IT and security requirements, while still having an up-to-date interoperable stack of SPHN-compliant terminologies. In the future, more terminologies and mappings will be added to the terminology service according to the needs of the SPHN community.

A Lightweight Private Internet of Things Remote System and its Micromixing Application

Most current Internet of Things systems face many problems like privacy leakage, information island and low application value, because they are designed from the perspective of Internet of Things service providers. However, the practicality and privacy security of Internet of Things services are more important from the perspective of Internet of Things system users. Considering that there aren’t too many Internet of Things devices to use and manage for ordinary people and service, a more service-oriented architecture of the lightweight private IOT remote system is proposed in this paper. To verify this system, verifications like remote sensors, remote actuators, and a micromixing application are developed and illustrated. Owing to its distributed server architecture, this system has high flexibility, practicability, parallelism, and low coupling. This novel system architecture provides a new perspective to develop practical valuable private Internet of Things services.

Pensando Distributed Services Architecture

The Pensando distributed services architecture is designed to provide programmable security, network, storage, and visibility services to a data center landscape growing in scale and complexity. This domain-specific architecture combines P4 packet processing pipelines with embedded ARM CPUs and hardware offload engines in a tightly coupled, cache coherent interconnect. Millions of flows are tracked by hardware and distributed to a combination of P4 programs, ARM processes, custom hardware engines, and PCIe based hosts and devices. There are currently 16- and 7-nm ASIC implementations of this distributed services architecture, future generations are in design.

Decomposition Techniques in Distributed Service Architecture

Decomposition Techniques in Distributed Service Architecture

Large-scale distributed semantic augmented reality services – A performance evaluation

This paper presents an architecture for large-scale distributed AR services and an experimental evaluation of the main component – Semantic Augmented Reality Middleware. The architecture is based on a client-server design, which supports semantic modeling and building contextual AR presentations for a large number of users. The architecture is based on the Service Oriented Architecture (SOA) paradigm that enables building distributed systems that provide application functionality as services to either end-user applications or other services. The proposed architecture consists of two client-side components and multiple server-side distributed service applications – semantic middleware and AR service providers. The presented findings are promising and demonstrate that application of semantic web techniques can be a useful solution to implement large-scale contextual distributed AR environments.

MonALISA, an agent-based monitoring and control system for the LHC experiments

MonALISA, which stands for Monitoring Agents using a Large Integrated Services Architecture, has been developed over the last fifteen years by California Insitute of Technology (Caltech) and its partners with the support of the software and computing program of the CMS and ALICE experiments at the Large Hadron Collider (LHC). The framework is based on Dynamic Distributed Service Architecture and is able to provide complete system monitoring, performance metrics of applications, Jobs or services, system control and global optimization services for complex systems. A short overview and status of MonALISA is given in this paper.

Engineering of web services for internet of things applications

This paper addresses web services for Internet of Things (IoT) applications. The research focus is on IoT applications for mobile monitoring and third party's database access. Both web services design aspects and deployment issues are studied. By synthesis of functional abstraction for configuration management and performance monitoring, an approach to design RESTful web services for mobile monitoring is suggested. In addition to traditional operations supported by Database Management Systems, web services' operations for subscription management and notifications about occurrences of specific data-related events, as well as operations for trust and security management are proposed in the context of third party's database access. Web services deployment issues concern the access control aimed to prevent from overloading. A formal method for a model-based load evaluation of SOA (Service Oriented Architecture) application server is suggested. The method takes into account the distributed server architecture and the processing of heterogeneous traffic. The SOA application server's utilization is evaluated by simulation.

The Client Assignment Problem for Continuous Distributed Interactive Applications: Analysis, Algorithms, and Evaluation

Interactivity is a primary performance measure for distributed interactive applications (DIAs) that enable participants at different locations to interact with each other in real time. Wide geographical spreads of participants in large-scale DIAs necessitate distributed deployment of servers to improve interactivity. In a distributed server architecture, the interactivity performance depends on not only client-to-server network latencies but also interserver network latencies, as well as synchronization delays to meet the consistency and fairness requirements of DIAs. All of these factors are directly affected by how the clients are assigned to the servers. In this paper, we investigate the problem of effectively assigning clients to servers for maximizing the interactivity of DIAs. We focus on continuous DIAs that changes their states not only in response to user operations but also due to the passing of time. We analyze the minimum achievable interaction time for DIAs to preserve consistency and provide fairness among clients, and formulate the client assignment problem as a combinatorial optimization problem. We prove that this problem is NP-complete. Three heuristic assignment algorithms are proposed and their approximation ratios are theoretically analyzed. The performance of the algorithms is also experimentally evaluated using real Internet latency data. The experimental results show that our proposed Greedy Assignment and Distributed-Modify Assignment algorithms generally produce near optimal interactivity and significantly reduce the interaction time between clients compared to the intuitive algorithm that assigns each client to its nearest server.

Slice embedding solutions for distributed service architectures

Network virtualization provides a novel approach to running multiple concurrent virtual networks over a common physical network infrastructure. From a research perspective, this enables the networking community to concurrently experiment with new Internet architectures and protocols. From a market perspective, on the other hand, this paradigm is appealing as it enables infrastructure service providers to experiment with new business models that range from leasing virtual slices of their infrastructure to hosting multiple concurrent network services. In this article, we present the slice embedding problem and recent developments in the area. A slice is a set of virtual instances spanning a set of physical resources. The embedding problem consists of three main tasks: (1) resource discovery, which involves monitoring the state of the physical resources, (2) virtual network mapping, which involves matching users' requests with the available resources, and (3) allocation, which involves assigning the resources that match the users' requests. We also outline how these three tasks are tightly coupled, and how there exists a wide spectrum of solutions that either solve a particular task or jointly solve multiple tasks along with the interactions among them. To dissect the space of solutions, we introduce three main classification criteria, namely: (1) the type of constraints imposed by the user, (2) the type of dynamics considered in the embedding process, and (3) the allocation strategy adopted. Finally, we conclude with a few interesting research directions.

ODSA: Chord-Based Object Discovery Service Architecture for the Internet of Things

In the emerging Internet of Things (IOT), as a means to fulfill item-level lookup, lookup service or discovery service plays a critical role. However, existing lookup service and discovery service of IOT mainly rely on a centralized or a chain-style framework, thus have some drawbacks or bottlenecks to prevent them from being widely adopted, while the issue of locating hot-point resource has received much less attention, as well as the item-level lookup service is still missing. Therefore, we present object discovery service architecture (ODSA), a distributed serial number level object discovery service architecture using Chord for the IOT, give its two kinds of framework and some relevant mechanisms, and mainly focus on the Double-Ring architecture for enhancing the lookup efficiency and relieving the congestion of network by balancing the flows of query. Additionally, we analyze theoretically the node number constraint relationship between inside ring and outside ring. We conduct a simulation study based on P2PSim of MIT and our prototype to validate and evaluate the analysis and our solution. Results from the simulation validate the correctness of the theoretical analysis and the availability, scalability of our solution. The related comparisons and analysis show that our solution is available and outperforms substantially other approaches of current IOT lookup service.

Optimizing Client Assignment for Enhancing Interactivity in Distributed Interactive Applications

Distributed interactive applications (DIAs) are networked systems that allow multiple participants at different locations to interact with each other. Wide spreads of client locations in large-scale DIAs often require geographical distribution of servers to meet the latency requirements of the applications. In the distributed server architecture, the network latencies involved in the interactions between clients are directly affected by how the clients are assigned to the servers. In this paper, we focus on the problem of assigning clients to appropriate servers in DIAs to enhance their interactivity. We formulate the problem as a combinational optimization problem and prove that it is NP-complete. Then, we propose several heuristic algorithms for fast computation of good client assignments and theoretically analyze their approximation ratios. The proposed algorithms are also experimentally evaluated with real Internet latency data. The results show that the proposed algorithms are efficient and effective in reducing the interaction time between clients, and our proposed Distributed-Modify-Assignment adapts well to the dynamics of client participation and network conditions. For the special case of tree network topologies, we develop a polynomial-time algorithm to compute the optimal client assignment.

Effective distributed service architecture for ubiquitous video surveillance

Video surveillance systems are playing an important role to protect lives and assets of individuals, enterprises and governments. Due to the prevalence of wired and wireless access to Internet, it would be a trend to integrate present isolated video surveillance systems by applying distributed computing environment and to further gestate diversified multimedia intelligent surveillance (MIS) applications in ubiquity. In this paper, we propose a distributed and secure architecture for ubiquitous video surveillance (UVS) services over Internet and error-prone wireless networks with scalability, ubiquity and privacy. As cloud computing, users consume UVS related resources as a service and do not need to own the physical infrastructure, platform, or software. To protect the service privacy, preserve the service scalability and provide reliable UVS video streaming for end users, we apply the AES security mechanism, multicast overlay network and forward error correction (FEC), respectively. Different value-added services can be created and added to this architecture without introducing much traffic load and degrading service quality. Besides, we construct an experimental test-bed for UVS system with three kinds of services to detect fire and fall-incident features and record the captured video at the same time. Experimental results showed that the proposed distributed service architecture is effective and numbers of services on different multicast islands were successfully connected without influencing the playback quality. The average sending rate and the receiving rates of these services are quite similar, and the surveillance video is smoothly played.

A Novel Strategy for Load Balancing of Distributed Medical Applications

Current trends in medicine, specifically in the electronic handling of medical applications, ranging from digital imaging, paperless hospital administration and electronic medical records, telemedicine, to computer-aided diagnosis, creates a burden on the network. Distributed Service Architectures, such as Intelligent Network (IN), Telecommunication Information Networking Architecture (TINA) and Open Service Access (OSA), are able to meet this new challenge. Distribution enables computational tasks to be spread among multiple processors; hence, performance is an important issue. This paper proposes a novel approach in load balancing, the Random Sender Initiated Algorithm, for distribution of tasks among several nodes sharing the same computational object (CO) instances in Distributed Service Architectures. Simulations illustrate that the proposed algorithm produces better network performance than the benchmark load balancing algorithms-the Random Node Selection Algorithm and the Shortest Queue Algorithm, especially under medium and heavily loaded conditions.

Public Agility and Change in a Network Environment

Preparing for change is increasingly core business for governmental organizations. The networked society and the increasing connectedness of governmental organizations have as much impact on the complexity of the change process as the complexities of the corpus of law. Change is not only driven by changes in the law; changes in the organization’s environment often create a need to redesign business processes, reallocate roles and responsibilities, and reorder tasks. Moreover, preparations for change are not limited to the internal processes and systems of these organizations. Propagation of changes to network partners and redesign of network arrangements can be an enormous challenge. In the AGILE project, we develop a design method, distributed service architecture, and supporting tools that enable organizations - administrative and otherwise - to orchestrate their law-based services in a networked environment. This paper explains the Agile approach and describes some of its key principles.

A distributed server architecture supporting dynamic resource provisioning for BPM-oriented workflow management systems

Workflow management systems have been widely used in many business process management (BPM) applications. There are also a lot of companies offering commercial software solutions for BPM. However, most of them adopt a simple client/server architecture with one single centralized workflow-management server only. As the number of incoming workflow requests increases, the single workflow-management server might become the performance bottleneck, leading to unacceptable response time. Development of parallel servers might be a possible solution. However, a parallel server architecture with a fixed-number of servers cannot efficiently utilize computing resources under time-varying system workloads. This paper presents a distributed workflow-management server architecture which adopts dynamic resource provisioning mechanisms to deal with the probable performance bottleneck. We implemented a prototype system of the proposed architecture based on a commercial workflow management system, Agentflow. A series of experiments were conducted on the prototype system for performance evaluation. The experimental results indicate that the proposed architecture can deliver scalable performance and effectively maintain stable request response time under a wide range of incoming workflow request workloads.