• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.42 no.5
• /
• pp.975-981
• /
• 2017

In this paper, we propose a method for efficient use of energy when using flash device in WSN environment. Typical Flash devices have a drawback to be an energy efficient storage media in the energy-constrained WSNs due to the high standby energy. An energy efficient approach to deploy Flash devices into WSNs is simply turning the Flash device off whenever idle. In this regard, we make the simple but ideal approach realistic by removing these two obstacles by exploiting nonvolatile RAM (NVRAM), which is an emerging memory technology that provides both non-volatility and byte-addressability. Specifically, we make use of NVRAM as an extension of metadata storage to remove the FTL metadata scanning process that mainly incurs the two obstacles. Through the implementation and evaluation in a real system environment, we verify that significant energy savings without sacrificing I/O performance are feasible in WSNs by turning off the Flash device exploiting NVRAM whenever it becomes idle. Experimental results show that the proposed method consumes only about 1.087% energy compared to the conventional storage device.

• Journal of Internet Computing and Services
• /
• v.17 no.2
• /
• pp.1-9
• /
• 2016

Users are provided infrastructure such as CPU, memory, network, and storage as IaaS (Infrastructure as a Service) service on cloud computing environments. However storage instances cannot support the maximum storage capacity that SQL servers can use, because the capacity of instances provided by service providers is usually limited. In this paper, we propose a method of securing mass storage capacity for SQL servers by sharing network disks with limited storage capacity. We confirmed through experiments that it is possible to secure mass storage capacity, which exceeds the maximum storage capacity provided by an instance with Amazon EBS on Amazon EC2 Windows environments, and it is possible to improve the overall performance of the SQL servers by increasing the disk capacity and performance.

• The Korean Journal of Applied Statistics
• /
• v.33 no.5
• /
• pp.569-578
• /
• 2020

By estimating conditional quantile functions of the response, quantile regression (QR) can provide comprehensive information of the relationship between the response and the predictors. In addition, kernel quantile regression (KQR) estimates a nonlinear conditional quantile function in reproducing kernel Hilbert spaces generated by a positive definite kernel function. However, it is infeasible to use the KQR in analysing a massive data due to the limitations of computer primary memory. We propose a divide and conquer based KQR (DC-KQR) method to overcome such a limitation. The proposed DC-KQR divides the entire data into a few subsets, then applies the KQR onto each subsets and derives a final estimator by aggregating all results from subsets. Simulation studies are presented to demonstrate the satisfactory performance of the proposed method.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2011.08a
• /
• pp.391-391
• /
• 2011

최근 나노에 대한 연구가 활성화되고 나노입자가 가지는 특성이 부각되면서 이를 소자 제조에 응용하고자 하는 연구가 집중적으로 이루어지고 있다. 박막에 포함된 나노입자는 메모리, 고효율 박막형 태양전지 등에 이용될 수 있는 가능성을 보여주었으며, 나노입자를 바탕으로 소자 제조에 관한 연구가 이루어지면서 플라즈마 내 발생하는 나노입자를 이용하여 패터닝 등에 적용하고자 하는 연구가 국내외에서 활발히 이루어지고 있다. 특히 플라즈마에서 발생하는 나노입자는 플라즈마 내 전기적 및 화학적 특징으로 인해 다른 입자 제조 공정과 달리 응집이 없는 균일한 입자를 제조할 수 있다. 이러한 플라즈마 내 발생 입자를 응용하기 위해서는 공정 조건에 따른 입자의 생성 및 성장 분석이 필요하다. 하지만 이러한 입자 발생 특성에 관한 연구는 기존에 밝혀진 반응 메커니즘으로 인해 수치해석적 연구는 체계적으로 진행되었으나 실험적 연구의 경우 적합한 측정 장비의 부재로 인해 제한이 있었다. 따라서 본 연구에서는 저압에서 실시간으로 나노입자 분포를 측정할 수 있는 PBMS (particle beam mass spectrometer)를 이용하여 나노입자 합성 공정 중 발생하는 입자의 존재를 확인하고 특성을 분석하였다. 실리콘 나노 입자의 측정은 PBMS 장비의 전단 부분을 PECVD (plasma enhanced chemical vapor deposition) 장치 내부에 연결하여 진행하였다. PECVD를 이용한 실리콘 나노입자 형성의 주요 변수는 RF pulse, 가스(Ar, SiH4, H2)의 유량, Plasma power, 공정압력 등이 있다. 본 연구에서는 실리콘 나노입자를 만드는데 필요한 여러 변수들을 제어함으로써 이에 따른 입경분포를 측정하였다. 또한 동일한 조건에서 생성 나노입자를 포집하여 TEM과 SEM을 이용하여 분석하여 그 결과를 비교하였다. 추후 지속적 연구에 의해 변수에 따른 나노입자 생성을 데이터베이스화 하여 요구되는 응용분야에 적합한 특성을 가지는 나노입자를 형성하는 조건을 정립 하는데 중요한 역할을 할 것을 기대할 수 있다.

• Journal of KIISE:Computing Practices and Letters
• /
• v.5 no.4
• /
• pp.395-404
• /
• 1999

포인터 스위즐링 기법은 포인터 스위즐링과 언스위즐링으로 이루어지며, 포인터 스위즐링은 객체 접근 시 객체 식별자를 해당 객체의 메모리 주소로 교체하는 것을 말하며, 언스위즐링은 객체 교체 또는 객체 저장 시에 스위즐링된 포인터를 원래의 객체 식별자로 환원하는 것을 말한다. 본 연구에서는 시스템 버퍼 구조에 따라 여러 포인터 스위즐링 기법을 분류하여 장단점을 분석하였으며, 이중 버퍼 구조 상에서 적극/소극, 직접/간접 스위즐링, 언스위즐링 모듈을 설계, 구현하였다. 또한 제한된 크기의 객체 버퍼 상에서 각 포인터 스위즐링 모듈의 성능을 평가하였다. 이 성능 평가의 결과로는 사용하지 않는 포인터는 스위즐링하지 않으며 언스위즐링 부담이 적은 소극 간접 스위즐링 기법이 일반적으로 뛰어난 성능을 나타냄을 알 수 있게 되었다.Abstract The pointer swizzling methods consist of pointer swizzling and unswizzling. Pointer swizzling replaces the OID of a object to the memory address of the object at object access time and unswizzling replaces the swizzling pointer of the OID at object replacement time or object save time. In this research, the different techniques for pointer swizzling are classified according to the system buffer structure and analyzed the pros and cons. In addition, eager/lazy, direct/indirect swizzling, unswizzling modules are designed and implemented on a dual buffering structure. Also, we evaluate the performance of pointer swizzling modules on the restricted object buffer size. The results of performance evaluation show that the performance of lazy indirect pointer swizzling technique is generally good because unused pointers are not swizzled, and unswizzling overhead is minimized.

• Journal of Internet Computing and Services
• /
• v.21 no.5
• /
• pp.149-157
• /
• 2020

Recently, as interest in artificial intelligence has increased, many studies have been conducted to implement AI processors. However, the AI processor requires functional verification as well as performance verification on whether the AI processor is suitable for the application. In this paper, We propose an AI processor performance analyzer that can verify the application performance and explore the limitations of the processor. By Using the performance analyzer, we explore the limitations of the AI processor and optimize the AI model to fit an AI processor in image recognition and speech recognition applications.

• Journal of Korea Society of Industrial Information Systems
• /
• v.17 no.6
• /
• pp.11-17
• /
• 2012

The HC-128 stram cipher which selected for the final eSTREAM portfolio is suitable for mobile Ad-Hoc network environments because of the ability of high-speed encryption in restricted memory space. In this paper, we analyzed the vulnerability of side channel analysis attack on HC-128 stream cipher. At the first, we explain a flaw of previous theoretical analysis result which defined the complexity of side-channel attack of HC-128 stream cipher as 'low' and then re-evaluate the security against side-channel attack by estimating the concrete complexity for recovering the secret key. As a result, HC-128 stream cipher is relatively secure against side-channel attack since recovering the secret key have $2^{65}$ computation complexity which is higher than other stream cipher's one.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.1B
• /
• pp.59-66
• /
• 2012

The use of MPSoC has been increasing because of a rise of use of mobile devices and complex applications. For improving the performance of MPSoC, number of processor has been increasing. Standard MPI is used for efficiently sending data in distributed memory architecture that has advantage in multi processor. Standard In this paper, we propose a scalable distributed memory system with a low cost hardware message passing interface(MPI). The proposed architecture improves transfer rate with buffered send for small size packet. Three queues, Ready Queue, Request Queue, and Reservation Queue, work as previous architecture, and two queues, Small Ready Queue and Small Request Queue, are added to send small size packet. When the critical point is set 8 bytes, the proposed architecture takes more than 2 times the performance improvement in the data that below the critical point.

• Journal of the Korea Society of Computer and Information
• /
• v.8 no.1
• /
• pp.51-59
• /
• 2003

It is effective to be located in the double nodes in the distributed object replication systems, then object which nodes share is the same contents. The nodes store an access information on their local cache as it access to the system. and then the nodes fetch and use it, when it needed. But with time the coherence Problems will happen because a data carl be updated by other nodes. So keeping the coherence of the system we need a mechanism that we managed the to improve to improve the performance and availability of the system effectively. In this paper to keep coherence in the shared memory condition, we can set the limited parallel performance without the additional cost except the coherence cost using it to keep the object at the proposed adaptive duplication object(ADO) algorithms. Also to minimize the coherence maintenance cost which is the bi99est overhead in the duplication method, we must manage the object effectively for the number of replication and location of the object replica which is the most important points, and then it determines the cos. And that we must study the adaptive duplication object management mechanism which will improve the entire run time.

• The Journal of the Korea Contents Association
• /
• v.8 no.9
• /
• pp.42-54
• /
• 2008

In recent years, embedded systems have been expanding their application domains from traditional embedded systems such as military weapons, robots, satellites and digital convergence systems such as celluar phones, PMP(Portable Multimedia Player), PDAs(Personal Digital Assistants) to Next Generation Personal Computers(NGPCs) such as eating PCs, wearable computers. The NGPCs are network-based, human-centric digital information devices diverged from the traditional PCs used mainly for document writing, internet searching and database management. Wearable computers with battery capacity and memory size limitations have to use real-time operating systems with small footprints and low power management techniques to provide user's QoS in spite of hardware constraints. In this paper, we have designed and implemented a low-power RTOS (called eRTOS) for wearable computers. The implemented eRTOS has 18KB footprints and the dynamic power management and the device power management schemes are adapted in it. Experimental results with wearable computer applications show that the low power techniques could save energy up to 47 %.