• KNOM Review
• /
• v.23 no.1
• /
• pp.43-50
• /
• 2020

As General Purpose Graphics Processing Unit (GPGPU) recently plays an essential role in high-performance computing, several cloud service providers offer GPU service. Most cluster orchestration platforms in a cloud environment using containers allocate the integer number of GPU to jobs and do not allow a node shared with other jobs. In this case, resource utilization of a GPU node might be low if a job does not intensively require either many cores or large size of memory in GPU. GPU virtualization brings opportunities to realize kernel concurrency and share resources. However, performance may vary depending on characteristics of applications running concurrently and interference among them due to resource contention on a node. This paper proposes GPU container co-execution framework with multiple server creation and execution based on Kubernetes, container orchestration platform for measuring interference which may be occurred by sharing GPU resources. Performance changes according to scheduling policies were investigated by executing several jobs on GPU. The result shows that optimal scheduling is not possible only considering GPU memory and computing resource usage. Interference caused by co-execution among applications is measured using the framework.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.3
• /
• pp.193-204
• /
• 2006

In most existing distributed Web systems, incoming requests are distributed to servers via Domain Name System (DNS). Although such systems are simple to implement, the address caching mechanism easily results in load unbalancing among servers. Moreover, modification of the DNS is necessary to load considering the server's state. In this paper, we propose a new dynamic load balancing method using dynamic DNS update and round-robin mechanism. The proposed method performs effective load balancing without modification of the DNS. In this method, a server can dynamically be added to or removed from the DNS list according to the server's load. By removing the overloaded server from the DNS list, the response time becomes faster. For dynamic scheduling, we propose a scheduling algorithm that considers the CPU, memory, and network usage. We can select a scheduling policy based on resources usage. The proposed system can easily be managed by a GUI-based management tool. Experiments show that modules implemented in this paper have low impact on the proposed system. Furthermore, experiments show that both the response time and the file transfer rate of the proposed system are faster than those of a pure Round-Robin DNS.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.3
• /
• pp.369-378
• /
• 2020

Due to the widespread use of mobile devices, smartphone penetration and usage rate continue to increase and there is also an increasing amount of data that need to be stored and managed in applications. Therefore, recent applications use mobile databases to store and manage user data. Realm database, developed in 2014, is attracting more attention from developers because of advantages of continuous updating, high speed, low memory usage, simplicity and readability of the code. It also supports an encryption to provide confidentiality and integrity of personal information stored in the database. However, since the encryption can be used as an anti-forensic technique, it is necessary to analyze the encryption and decryption processes provided by Realm Database. In this paper, we analyze the structure of Realm Database and its encryption and decryption process in detail, and analyze an application that supports an encryption to propose the use cases of the Realm Database.

• The KIPS Transactions:PartC
• /
• v.13C no.7 s.110
• /
• pp.821-830
• /
• 2006

As internet worms are spread out worldwide, the detection and filtering of worms becomes one of hot issues in the internet security. As one of implementation methods to detect worms, the Linux Netfilter kernel module can be used. Its basic operation for worm detection is a string matching where coming packet(s) on the network is/are compared with predefined worm signatures(patterns). A worm can appear in a packet or in two (or more) succeeding packets where some part of worm is in the first packet and its remaining part is in its succeeding packet(s). Assuming that the maximum length of a worm pattern is less than 1024 bytes, we need to perform a string matching up to two succeeding packets of 2048 bytes. To do so, Linux Netfilter keeps the previous packet in buffer and performs matching with a combined 2048 byte string of the buffered packet and current packet. As the number of concurrent connections to be handled in the worm detection system increases, the total size of buffer (memory) increases and string matching speed becomes low In this paper, to reduce the memory buffer size and get higher speed of string matching, we propose a string matching scheme without using buffer. The proposed scheme keeps the partial matching result of the previous packet with signatures and has no buffering for previous packet. The partial matching information is used to detect a worm in the two succeeding packets. We implemented the proposed scheme by modifying the Linux Netfilter. Then we compared the modified Linux Netfilter module with the original Linux Netfilter module. Experimental results show that the proposed scheme has 25% lower memory usage and 54% higher speed compared to the original scheme.

• KIEAE Journal
• /
• v.15 no.6
• /
• pp.93-100
• /
• 2015

Purpose: Ballparks of KBO which were built by local governments and operated for baseball game-centered have been required more sustainable development according to going into low growth phase in Korea recently. MLB ballparks with the teams having 100 year old tradition have been sustainable-developed economically, socially, environmentally and culturally. This research is to study the characteristics of cultural sustainability in architectural design of 30 MLB parks. Method: To begin with comparison analysis of usage rate of 10 ballparks of KBO with 30 ballparks of MLB, and architectural designs of facades, fields, accommodations, sculptures, greens, roof gardens, etc. are analyzed in the MLB ballpark. And finally, the characteristics of cultural sustainability in the architectural design are analyzed. Result: MLB ballparks have played role as core-space of urban community, accumulated space of citizens' memory being originated in natural climatic feature of region, historical image of city and tradition of home-ballpark. A basis of these characteristics could is nature of cultural sustainability, that is to say local community, historical restoration, social solidarity.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.4 s.304
• /
• pp.153-158
• /
• 2005

In this paper, a filter is proposed to compensate droop of the CIC filter for SDR. The proposed compensation filter has almost linear-phase characteristic, requires low operational complexity, and is cost-effective due to its second-order characteristic and lowest operational rate in the baseband.. Especially, it compensates droop in the passband with little performance degradation in the stopband. It is shown, by a design example and its performance analysis, that the proposed compensation method gives performance enhancement in communication systems. It is also shown that the proposed method is superior to conventional ones in view of memory usage and computational load.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.06a
• /
• pp.157-159
• /
• 2014

In this paper, we introduce SlimSRP, an ultra-low-power digital signal processor (DSP) solution for mobile audio and voice applications. So far, application processors (APs) have taken charge of all the tasks in mobile devices. However, they have suffered from short battery life problems to deal with complex usage scenarios, such as always-on voice trigger with continuous audio playback. From extensive analysis of audio and voice application characteristics, SlimSRP is designed to relive the performance and power burden of APs. It employs three-issue VLIW architecture, and the major low-power and high-performance techniques include: (1) an optimized register-file architecture friendly for constants generation, (2) a powerful instruction set to reduce the number of register file accesses and (3) a unique instruction compression scheme that contributes to saved memory size and reduced cache miss. An implementation of SlimSRP runs at up to 200MHz and the logic occupies 95K NAND2 gates in Samsung 28LPP process. The experimental results demonstrate that a MP3 decoder application with a 128kbps 44.1kHz input can run at 5.1MHz and the logic consumes only 22uW/MHz.

• Journal of KIISE:Software and Applications
• /
• v.32 no.7
• /
• pp.625-635
• /
• 2005

In this paper, we propose a hybrid approach for segmenting the lungs efficiently and automatically in chest CT images. The proposed method consists of the following three steps. first, lungs and airways are extracted by two- and three-dimensional automatic seeded region growing and connected component labeling in low-resolution. Second, trachea and large airways are delineated from the lungs by two-dimensional morphological operations, and the left and right lungs are identified by connected component labeling in low-resolution. Third, smooth and accurate lung region borders are obtained by refinement based on image subtraction. In experiments, we evaluate our method in aspects of accuracy and efficiency using 10 chest CT images obtained from 5 patients. To evaluate the accuracy, we Present results comparing our automatic method to manually traced borders from radiologists. Experimental results show that proposed method which use connected component labeling in low-resolution reduce processing time by 31.4 seconds and maximum memory usage by 196.75 MB on average. Our method extracts lung surfaces efficiently and automatically without additional processing like hole-filling.

• Smart Media Journal
• /
• v.4 no.2
• /
• pp.62-67
• /
• 2015

During several decades, hard disk drive(HDD) has been used in most computer systems as secondary storage and, however, the performance enhancement of HDD is limited by its mechanical properties. On the other hand, although the flash memory based solid state drive (SSD) has more advantages over HDD such as high performance and low noise, SSD is still too expensive for common usage and expected to take several years to replace HDD completely. Therefore, SSD caching mechanism using the SSD as a cache of high capacity HDD has been highlighted lately. The representatives of SSD caching mechanisms are typically bcache, dm-cache, Flashcache, and EnhanceIO. Each of them has its own internal mechanism and implementation, and this makes them to show their own pros. and cons. In this paper, we analyze the characteristics of each SSD caching mechanisms and compare the performance of them under various workloads. We expect that our contribution will be useful to enhance the performance of SSD caching mechanisms.

• Convergence Security Journal
• /
• v.12 no.4
• /
• pp.9-15
• /
• 2012

In recently years, repackaged malwares are becoming increased rapidly in Android smartphones. The repackaging is a technique to disassemble an app in a market, modify its source code, and then re-assemble the code, so that it is commonly used to make malwares by inserting malicious code in an app. However, it is impossible to collect all the apps in many android markets including too many apps. To solve the problem, we propose RePAD (RePackaged App Detector) scheme that is composed of a client and a remote server. In the smartphone-side, the client extracts the information of an app with low CPU overhead when a user installs the app. The remote server analyzes the information to decide whether the app is repackaged or not. Thus, the scheme reduces the time and cost to decide whether apps are repackaged. For the experiments, the client and server are implemented as an app on Galaxy TAB and PC respectively. We indicated that seven pairs of apps among ones collected in official and unofficial market are repackaged. Furthermore, RePAD only increases the average of CPU overhead of 1.9% and the maximum memory usage of 3.5 MB in Galaxy TAB.