Search Results (9)

Deep learning is a growing technique used to solve complex artificial intelligence (AI) problems. Large-scale deep learning has become a significant issue as a result of the expansion of datasets and the complexity of deep learning models. For training large-scale models, the cloud can be used as a distributed HPC (high-performance computing) tool with benefits in cost and flexibility. However, one of the major performance barriers in distributed deep learning in a distributed HPC environment is the network. The performance is often limited by heavy traffic like many stochastic gradient descent transfers for distributed communication. There are many network studies in distributed deep learning to solve these problems, but most research only focuses on improving communication performance and applying new methods or algorithms like overlapping parameter synchronization to minimize communication delay rather than considering the actual network. In this paper, we are focusing on the actual network, especially in a distributed HPC environment. In such an environment, if cluster nodes are assigned to different zones/regions which means a set of an appropriate number of distributed HPC nodes when performing distributed deep learning tasks, performance degradation due to network delay may occur. The proposed network optimization algorithm ensures that distributed work is placed in the same zone as much as possible to reduce network delay. Furthermore, scoring using network monitoring tools like loss, delay, and throughput is applied to select the optimal node within the zone. Our proposal has been validated on the Kubernetes platform, an open source orchestrator for the automatic management and deployment of micro-services. The performance of distributed deep learning is improved through the proposed scheduler. Full article

Here, we introduce the current stage and future directions of the wireless infrastructure of the Korea Research Environment Open NETwork (KREONET), a representative national research and education network in Korea. In 2018, ScienceLoRa, a pioneering wireless network infrastructure for scientific applications based on low-power wide-area network technology, was launched. Existing in-service applications in monitoring regions, research facilities, and universities prove the effectiveness of using wireless infrastructure in scientific areas. Furthermore, to support the more stringent requirements of various scientific scenarios, ScienceLoRa is evolving toward ScienceIoT by employing high-performance wireless technology and distributed computing capability. Specifically, by accommodating a private 5G network and an integrated edge computing platform, ScienceIoT is expected to support cutting-edge scientific applications requiring high-throughput and distributed data processing. Full article

As many industrial applications require real-time and reliability communication, a variety of routing graph construction schemes were proposed to satisfy the requirements in Industrial Wireless Sensor Networks (IWSNs). Each device transmits packet through a route which is designated based on the graph. However, as existing studies consider a network consists of static devices only, they cannot cope with the network changes by movement of mobile devices considered important in the recent industrial environment. Thus, the communication requirements cannot be guaranteed because the existing path is broken by the varying network topology. The communication failure could cause critical problems such as malfunctioning equipment. The problem is caused repeatedly by continuous movement of mobile devices, even if a new graph is reconstructed for responding the changed topology. To support mobile devices exploited in various industrial environments, we propose a Hierarchical Routing Graph Construction (HRGC). The HRGC is consisted of two phases for hierarchical graph construction: In first phase, a robust graph called skeleton graph consisting only of static devices is constructed. The skeleton graph is not affected by network topology changes and does not suffer from packet loss. In second phase, the mobile devices are grafted into the skeleton graph for seamless communication. Through the grafting process, the routes are established in advance for mobile device to communicate with nearby static devices in anywhere. The simulation results show that the packet delivery ratio is improved when the graph is constructed through the HRGC. Full article

In wireless sensor networks (WSNs), detection and report of continuous object, such as forest fire and toxic gas leakage, is one of the major applications. In large-scale continuous object tracking in WSNs, there might be many source nodes simultaneously, detecting the continuous object. Each nodes reports its data to both a base station and mobile workers in the industry field. For communication between the source nodes and a mobile worker, sink location service is needed to continuously notify the location of the mobile worker. But, as the application has a large number of sources, it causes a waste of energy consumption. To address this issue, in this paper, we propose a two-phase sink location service scheme. In the first phase, the proposed scheme constructs a virtual grid structure for merging the source nodes. Then, the proposed scheme aggregates the merging points from an originated merging point as the second phase. Simulation results show that the proposed scheme is superior to other schemes in terms of energy consumption. Full article

Mobile Cyber-Physical Systems (MCPS) have extended the application domains by exploiting the advantages of Cyber-Physical Systems (CPS) through the mobile devices. The cooperation of various mobile equipment and workers based on the MCPS further improved efficiency and productivity in the industry. To support this cooperation of groups of workers (hereafter referred to as the Mobile Sink Groups), data should be delivered to appropriate groups of workers in a timely manner. Traditionally, the data dissemination for MSG relies on flooding-based geocasting into the movable area of the group due to frequent movements of each group member. However, the flooding-based data dissemination could not be directly applied to real-time data delivery that demands the required time deadline and the end-to-end delivery distance, because the flooding could not define the end-to-end distance and progress to each member in a group. This paper proposes a real-time data delivery mechanism for supporting MSG in time-critical applications. In our mechanism, a ring-based modeling and data transfer scheme on a virtual grid in the ring for group mobility provides the end-to-end distance and the progress to forward real-time data to each member. Simulation results show our mechanism is superior to the existing ones in terms of real-time communication for MSG. Full article

Recent development of distributed cloud environments requires advanced network infrastructure in order to facilitate network automation, virtualization, high performance data transfer, and secured access of end-to-end resources across regional boundaries. In order to meet these innovative cloud networking requirements, software-defined wide area network (SD-WAN) is primarily demanded to converge distributed cloud resources (e.g., virtual machines (VMs)) in a programmable and intelligent manner over distant networks. Therefore, this paper proposes a logically isolated networking scheme designed to integrate distributed cloud resources to dynamic and on-demand virtual networking over SD-WAN. The performance evaluation and experimental results of the proposed scheme indicate that virtual network convergence time is minimized in two different network models such as: (1) an operating OpenFlow-oriented SD-WAN infrastructure (KREONET-S) which is deployed on the advanced national research network in Korea, and (2) Mininet-based experimental and emulated networks. Full article

As the Internet of Things (IoT) has developed, the emerging sensor network (ESN) that integrates emerging technologies, such as autonomous driving, cyber-physical systems, mobile nodes, and existing sensor networks has been in the limelight. Smart homes have been researched and developed by various companies and organizations. Emerging sensor networks have some issues of providing secure service according to a new environment, such as a smart home, and the problems of low power and low-computing capacity for the sensor that previous sensor networks were equipped with. This study classifies various sensors used in smart homes into three classes and contains the hierarchical topology for efficient communication. In addition, a scheme for establishing secure communication among sensors based on physical unclonable functions (PUFs) that cannot be physically cloned is suggested in regard to the sensor’s low performance. In addition, we analyzed this scheme by conducting security and performance evaluations proving to constitute secure channels while consuming fewer resources. We believe that our scheme can provide secure communication by using fewer resources in a smart home environment in the future. Full article

The darknet (i.e., a set of unused IP addresses) is a very useful solution for observing the global trends of cyber threats and analyzing attack activities on the Internet. Since the darknet is not connected with real systems, in most cases, the incoming packets on the darknet (‘the darknet traffic’) do not contain a payload. This means that we are unable to get real malware from the darknet traffic. This situation makes it difficult for security experts (e.g., academic researchers, engineers, operators, etc.) to identify whether the source hosts of the darknet traffic are infected by real malware or not. In this paper, we present the overall procedure of the in-depth analysis between the darknet traffic and IDS alerts using real data collected at the Science and Technology Cyber Security Center (S&T CSC) in Korea and provide the detailed in-depth analysis results. The ultimate goal of this paper is to provide practical experience, insight and know-how to security experts so that they are able to identify and trace the root cause of the darknet traffic. The experimental results show that correlation analysis between the darknet traffic and IDS alerts is very useful to discover potential attack hosts, especially internal hosts, and to find out what kinds of malware infected them. Full article

Legacy networks do not open the precise information of the network domain because of scalability, management and commercial reasons, and it is very hard to compute an optimal path to the destination. According to today’s ICT environment change, in order to meet the new network requirements, the concept of software-defined networking (SDN) has been developed as a technological alternative to overcome the limitations of the legacy network structure and to introduce innovative concepts. The purpose of this paper is to propose the application that calculates the optimal paths for general data transmission and real-time audio/video transmission, which consist of the major services of the National Research & Education Network (NREN) in the SDN environment. The proposed SDN routing computation (SRC) application is designed and applied in a multi-domain network for the efficient use of resources, selection of the optimal path between the multi-domains and optimal establishment of end-to-end connections. Full article