• Journal of Korea Society of Industrial Information Systems
• /
• v.13 no.1
• /
• pp.46-56
• /
• 2008

Double Patching is a multicasting technique for a VoD system which has been proposed to provide a true VoD service by making clients share a long patching stream as well as a regular stream For subsequent short patching streams, the technique always makes the long patching stream have extra data that will be played back during a double period of a patching window. In this paper, we propose a technique, using the start time of the latest short patching stream, optimizes Double Patching by deleting the useless data included in the long patching stream when the patching window of the long patching stream closes. The mean requirement for the server's bandwidth to provide the true VoD service is used as a performance metric, and the effect of the request inter-arrival time, the size of the client's local buffer and the video length on the mean bandwidth requirement is evaluated. Performance evaluation result shows that the proposed technique optimizes Double Patching in all cases.

• Journal of KIISE:Computing Practices and Letters
• /
• v.6 no.1
• /
• pp.21-35
• /
• 2000

MIDAS-II is the storage system for BADA DBMS developed at ETRI. This paper describes the extension of MIDAS-II for incorporating the tertiary storage device such as an optical disk jukebox or a tape library, enabling MIDAS-II to function as a storage system of the data server that stores a massive amount of multimedia data. The MIDAS-II disk volume structure is extended to efficiently function as a volume for the tertiary storage device with multiple platters, which canstore huge amount of data of the order of tera bytes. The storage structure of the LOB is changed to efficiently manage the LOB data in the tertiary storage device. The data structures of the shared memory, the process structure, and the utilities in MIDAS-II are also extended to efficiently incorporating the tertiary storage device. The functionalities of each MIDAS-II API function are expanded to handle the tertiary storage device, while the prototypes of those functions are intact in order not to affect the existing application programs. The performance evaluation shows that the extended MIDAS-II works effectively with the tertiary storage device. All these extensions and the performance evaluation are conducted in the SunOS 5.4 environment.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.12
• /
• pp.497-504
• /
• 2018

This paper studies about the Oriental Medical Expert System, based on Open Source Drools for rule engine processing, which contains scalability, availability, and modifiability. The system is developed with the Spring MVC framework and Ajax for stable services of the Web-based Medical Expert System. The diagnosis and treatment process of this Medical Expert system provides a service that provides the general users to accesses the web with a series of questionnaires. In order to compensate for the asynchronous communication between clients and services, and also for the complicated JDBC weaknesses, we applied the data handling in JSON to reduce the servers' loads, and also the Mybatis framework to improve the performance of the RDBMS, respectively. In addition, as the number of users increases to cope with the maximum available services of the web-based system, the load balancing structure using Nginx has been developed to solve the server traffic problems and the service availability has been increased. The experimental results show the stable services by approving the scalability test.

• Asia-pacific Journal of Multimedia Services Convergent with Art, Humanities, and Sociology
• /
• v.9 no.1
• /
• pp.179-187
• /
• 2019

With the recent development of small computing devices, IoT sensor network can be widely deployed and is now readily available with sensing, calculation and communi-cation functions at low cost. Sensor data management is a major component of the Internet of Things environment. The huge volume of data produced and transmitted from sensing devices can provide a lot of useful information but is often considered the next big data for businesses. New column-wise compression technology is mounted to the large data server because of its superior space efficiency. Since sensor nodes have narrow bandwidth and fault-prone wireless channels, sensor-based storage systems are subject to incomplete data services. In this study, we will bring forth a short overview through providing an analysis on IoT sensor networks, and will propose a new storage management scheme for IoT data. Our management scheme is based on RAID storage model using column-wise segmentation and compression to improve space efficiency without sacrificing I/O performance. We conclude that proposed storage control scheme outperforms the previous RAID control by computer performance simulation.

• KIPS Transactions on Computer and Communication Systems
• /
• v.8 no.10
• /
• pp.231-238
• /
• 2019

Analyzing next-generation genome sequencing data in a conventional way using single server may take several tens of hours depending on the data size. However, in order to cope with emergency situations where the results need to be known within a few hours, it is required to improve the performance of a single genome analysis. In this paper, we propose a parallelized method for pre-processing genome sequence data which can reduce the analysis time by utilizing the big data technology and the highperformance computing cluster which is connected to the high-speed network and shares the parallel file system. For the reliability of analytical data, we have chosen a strategy to parallelize the existing analytical tools and algorithms to the new environment. Parallelized processing, data distribution, and parallel merging techniques have been developed and performance improvements have been confirmed through experiments.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.21 no.2
• /
• pp.119-129
• /
• 2011

TCP-related Flooding attacks still dominate Distributed Denial of Service Attack. It is a great challenge to accurately detect the TCP flood attack in hish speed network. In this paper, we propose the NIC_Cookie logic implementation, which is a kind of security offload engine against TCP-related DDoS attacks, on network interface card. NIC_Cookie has robustness against DDoS attack itself and it is independent on server OS and external network configuration. It supports not IP-based response method but packet-level response, therefore it can handle attacks of NAT-based user group. We evaluate that the latency time of NIC_Cookie logics is $7{\times}10^{-6}$ seconds and we show 2Gbps wire-speed performance through a benchmark test.

• Journal of Internet of Things and Convergence
• /
• v.7 no.2
• /
• pp.47-53
• /
• 2021

In order to transmit video data stored in servers with limited bandwidth in IoT environments to many clients, a transmission plan must be established by considering factors such as the number of transmission rate changes, peak transmission rate, and transmission rate changes. This transmission plan is called smoothing, and includes CBA that minimizes the number of transmission rate increases, MCBA that minimizes the number of transmission rate changes, and MVBA to minimize the transmission rate changes. In this work, to evaluate the performance of the proposed algorithm[16], we compare the proposed algorithm with the existing smoothing algorithms with the peak rate, the number of transmission rate changes, the rate increase, the peak rate utilization, and the average transmission rate with various video data and buffer sizes. The evaluation results show that the proposed algorithm helps to efficiently use the server's finite network resources by establishing a transport plan with the lowest average transfer rate.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.6
• /
• pp.1045-1054
• /
• 2020

In this paper, we designed a TCP/IP based ring buffer system that can stably transfer bulk data streams in the unstable network environments. In the scheme we proposed, The observation data stream generated and output by each radio observatory's backend system as a UDP frame is stored as a UDP packet in a large capacity ring buffer via a socket buffer in the client system. Thereafter, for stable transmission to the remote destination, the packets are processed in TCP and transmitted to the socket buffer of server system in the correlation center, which packets are stored in a large capacity ring buffer if there is no problem with the packets. In case of errors such as loss, duplication, and out of order delivery, the packets are retransmitted through TCP flow control, and we guaranteed that the reliability of data arriving at the correlation center. When congestion avoidance occurs due to network performance instability, we also suggest that performance degradation can be minimized by applying parallel streams.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.3
• /
• pp.469-478
• /
• 2022

This paper contains the results of developing smart devices and applications to monitor the load power of the industrial manufacturing machine and evaluate its performance. The smart devices in this paper are divided into two functionalities, which are collecting load power along with operating environment data of industrial manufacturing machines and transmitting the data to servers. Load power data collected from the smart devices are uploaded to MariaDB inside the Amazon Web Service (AWS) server. Using the RESTFul API, the uploaded power data can be retrieved and shown on the web and mobile application in the form of a graph to provide monitoring capability. To evaluate the performance of the developed system, the response time from MariaDB to web and mobile applications was measured. The results is ranging from 0.0256 to 0.0545 seconds in a 4G (LTE) network environment and from 0.6126 to 1.2978 seconds in a 3G network environment, which is considered a satisfactory result.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.7
• /
• pp.205-216
• /
• 2022

One of the many tools used for distributed deep learning training is Kubeflow, which runs on Kubernetes, a container orchestration tool. TensorFlow jobs can be managed using the existing operator provided by Kubeflow. However, when considering the distributed deep learning training jobs based on the parameter server architecture, the scheduling policy used by the existing operator does not consider the task affinity of the distributed training job and does not provide the ability to dynamically allocate or release resources. This can lead to long job completion time and low resource utilization rate. Therefore, in this paper we proposes a new operator that efficiently schedules distributed deep learning training jobs to minimize the job completion time and increase resource utilization rate. We implemented the new operator by modifying the existing operator and conducted experiments to evaluate its performance. The experiment results showed that our scheduling policy improved the average job completion time reduction rate of up to 84% and average CPU utilization increase rate of up to 92%.