• Broadcasting and Media Magazine
• /
• v.22 no.4
• /
• pp.11-17
• /
• 2017

본 고에서는 멀티 응용 데이터 플랫폼으로 진화하고 있는 하둡(Hadoop) 2.0의 주요 특징과 관련된 다양한 데이터 처리 프레임워크들에 대해서 기술하고자 한다. 기존의 맵리듀스(MapReduce) 기반의 배치 처리(Batch Processing)에 최적화되어 있던 하둡 1.0과는 달리, YARN의 등장과 함께 시작된 하둡 2.0 플랫폼은 다양한 형태의 데이터 처리 워크플로우들(Batch, Interactive, Streaming 등)을 지원할 수 있는 기능을 제공하고 있다. 또한, 최근에는 고성능컴퓨팅 분야에서 주로 활용되던 기술들도 하둡 2.0 플랫폼에서 지원되고 있다. 마지막으로 YARN 어플리케이션 개발 사례로서 본 연구팀에서 개발 중에 있는 Many-Task Computing (MTC) 응용을 위한 신규 데이터 처리 프레임워크를 소개한다.

• Journal of Microbiology and Biotechnology
• /
• v.9 no.2
• /
• pp.179-183
• /
• 1999

A recombinant form of hirudin, a potent thrombin-specific inhibitor derived from the bloodsucking leech, was expressed as a secretory product in Saccharomyces cerevisiae under the control of GALl0 promoter and the mating factor $\alpha$pre-pro leader sequence. In an attempt to produce recombinant hirudin (r-Hir) of therapeutic purity in large quantities, the fed-batch fermentation was carried out by using this recombinant yeast, and subsequently downstream processing was developed with the preparative-scale column chromatography systems. About 234 mg/l of biologically active r-Hir was produced as a secretory product by the fed-batch fermentation strategy developed for an efficient downstream processing. Using a two-step chromatography process (an anion exchange chromatography followed by the reverse phase HPLC), the r-Hir was purified to>98% with an overall recovery yield of 84%. According to the N-terminal amino acid sequencing, the purified r-Hir was found to have the predicted N-terminal amino acid sequence. The biological activity of the purified r-Hir to inhibit thrombin was also identical to that of the commercial hirudin.

• Journal of KIISE:Computing Practices and Letters
• /
• v.5 no.5
• /
• pp.583-594
• /
• 1999

With the advance of VLSI and network technologies, it has now become a common practice to deploy a various forms of distributed computing environments. A study shows that a lot of network-aware computers are in an idle state for considerable amount of time depending on the types of users and time frames of the day. If we can take the full advantage of those idle computers, we can obtain the enormous combined processing power without further costly investment. In this paper, we present a distributed batch-processing system, called the Java Distributed Batch-processing System (JDBS), which allows us to execute CPU-intensive, independent jobs across a pool of idle workstations on top of extant distributed computing environments. Since JDBS is implemented using a Java programming language, it not only extends the scope of machine types that can be joined to the pool, but makes it a lot easier to build an entire system. Besides, JDBS is scalable and fault-tolerant due to its multi-cluster organization and intelligent strategies. A graphical user interface is also provided to facilitate the registration and unregistration, job submission, and job monitoring.

• Proceedings of the KSR Conference
• /
• 2010.06a
• /
• pp.609-614
• /
• 2010

In this paper, the structural integrity and fatigue strength for the bogie frame of Monorail being developed in domestic was evaluated. Presently, the standard of evaluation for the bogie frame of monorail was not regulated. Therefore, the evaluation of the structural integrity and fatigue strength for the bogie frame was performed on the basis of the UIC 615-4 standard. The structural integrity of the designed bogie frame was evaluated by displacement and Von-Mises stress under each load conditions. And the fatigue strength was evaluated by combined main in-service load conditions specified at UIC 615-4 standard and it was compared with result of fatigue analysis using winLIFE v3.1 with the function of batch processing. The results shows that the structural integrity and fatigue strength of the designed bogie frame was satisfied, and the fatigue analysis using batch processing was more effective than conventional fatigue analysis using combined load conditions.

• Journal of the Korea Society for Simulation
• /
• v.14 no.1
• /
• pp.43-53
• /
• 2005

We developed the module of the software that robot designers can perform their work faster and more easily. The parametric modeler is founded on the virtual robot design program. The virtual robot design program is the powerful software which may be used to solve various problems of robot kinematics and dynamics. The parametric modeler in the software we developed is that all the positions of joints and links are changed automatically when the designer changes one joint or one link in the robot system. Without parametric method, robot-designers must change all the positions of connected joints and links. It might become time-consuming. However, it is very efficient for designers to use the method of batch-processing in performing design-changes of robot-links using the parametric modeler.

• International Journal of Computer Science & Network Security
• /
• v.21 no.11
• /
• pp.23-30
• /
• 2021

Sentiment Analysis has become very important field of research because posting of reviews is becoming a trend. Supervised, unsupervised and semi supervised machine learning methods done lot of work to mine this data. Feature engineering is complex and technical part of machine learning. Deep learning is a new trend, where this laborious work can be done automatically. Many researchers have done many works on Deep learning Convolutional Neural Network (CNN) and Long Shor Term Memory (LSTM) Neural Network. These requires high processing speed and memory. Here author suggested two models simple & bidirectional deep leaning, which can work on text data with normal processing speed. At end both models are compared and found bidirectional model is best, because simple model achieve 50% accuracy and bidirectional deep learning model achieve 99% accuracy on trained data while 78% accuracy on test data. But this is based on 10-epochs and 40-batch size. This accuracy can also be increased by making different attempts on epochs and batch size.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.05a
• /
• pp.60-62
• /
• 2020

데이터의 양은 기술의 발전으로 발생하는 크게 증가하였고 다양한 빅데이터 처리 플랫폼이 등장하고 있다. 이 중 가장 널리 사용되고 있는 품랫폼이 Apache 소프트웨어 재단에서 개발한 Hadoop이며, Hadoop은 IoT 분야에도 사용된다. 그러나 기존에 Hadoop 기반 IoT 센서 데이터 수집 분석 환경은 Hadoop의 코어 프로젝트인 HDFS의 Small File로 인한 네임노드의 과부하 문제와 Import된 데이터의 Update나 Delete가 불가능하다는 문제가 있다. 본 논문에서는 Apache Kudu와 Impala를 활용해 Lambda Architecture를 설계한다. 제안하는 Architecture는 IoT 센서 데이터를 Cold-Data와 Hot-Data로 분류해 각 성격에 맞는 스토리지에 저장하고 Batch를 동해 생성된 Batch-View와 Apache Kudu와 Impala를 통해 생성된 Real-time View를 활용해 기존 Hadoop 기반 IoT 센서 데이터 수집 분석 환경의 문제를 해결하고 사용자가 분석된 데이터에 접근하는 시간을 단축한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.2-4
• /
• 2024

한국과학기술정보연구원(이하 KISTI)의 국가슈퍼컴퓨팅센터에서는 슈퍼컴퓨터 5호기인 Nurion과 Neuron 시스템을 구축하여 국내 연구자들에게 서비스하고 있다. 이 중에서 Neuron 시스템은 GPU 클러스터 시스템으로 SLURM Batch Scheduler를 이용하여 공동활용서비스를 제공하고 있다. 본 논문에서는 Neuron에서 사용 중인 SLURM Batch Scheduler와 리눅스의 crontab 기능을 이용하여 소프트웨어 장애가 발생한 계산노드를 자동으로 복구시키는 기능을 구현하여 장애처리 대기시간을 단축시키는 기법에 대해서 소개한다.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.28 no.4
• /
• pp.105-114
• /
• 2003

This paper considers a single facility scheduling problem for the fabrication of components supporting the subsequent assembly into products. Each product consists of two types of components, one is common to all products and the other is unique to itself. The unique components are processed individually and the common components are processed in batches, and a distinct setup is incurred for the production of a batch of common components. Under the so-called “batch availability” environment, we propose an efficient dynamic programming(DP) procedure for the optimal sequencing and batching schedule with minimum total completion times of products in the situation that each product is said to be complete if both common and unique components of the product are finished. Computational results show that the proposed DB procedure is more efficient than the previous work which has considered the same problem as this paper.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.38 no.1
• /
• pp.61-67
• /
• 2013

We consider a problem of grouping orders for batch processing that arises in production systems where customer orders are processed in batches. This problem can be viewed as a variant of bin packing problem where items can be split and a pair of items can be placed in a bin when the items are compatible with each other. In this paper, we consider a special case that at most two different items can be placed in a single bin and the size of every item is at most the size of a bin.