• The KIPS Transactions:PartA
• /
• v.17A no.6
• /
• pp.259-264
• /
• 2010

In case of scheduling tasks with shared resources in multiprocessor systems, Global Earliest Deadline First (GEDF) algorithm, equally applied Earliest Deadline First (EDF) which runs scheduling with deadline criterion, makes schedulability decline because GEDF typically does not have a specific process in order to handle tasks with shared resources. In this paper, we propose Earliest Deadline First with Partitioning (EDFP) for tasks with shared resources which partitions a task into two kinds of subtasks that include critical sections to access to shared resources, gives their own deadline respectively and manages them. As a result of simulations, EDFP shows better performance than GEDF for tasks with shared resources since system load goes up and the number of processor increases.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.5
• /
• pp.1421-1444
• /
• 2012

Mobile devices such as Android devices are emerging as a convenient client computing device with mobility and context-sensing capability. However, the computing power and hardware resource of the devices are limited due to their small form-factor. Consequently, large-scaled applications could not be deployed on these devices. Nonetheless, if the large-scaled applications are deployed and executed on the devices, high performance of the applications cannot be guaranteed. To remedy the limitation in terms of performance, it is inevitable to let some heavy-weight functionality executed on the server side and let a client application invoke the functionality in the server. To realize this kind of mobile applications, we adopt well-defined architecture design principles; being thin-client, being layered with Model-View-Controller (MVC), and being balanced between client side and server side. By adopting the principles, we propose a unique, ideal and practical architecture for mobile applications, called balanced MVC architecture. By considering the principles, key design considerations of realizing balanced MVC architecture lie in functionality partitioning. Hence, we define key criteria of determining the degree of performance. And, we define a method to design a balanced MVC architecture which embodies functionality partitioning for high performance, and a simulation-based evaluation method of balanced MVC architectures.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.32 no.3
• /
• pp.487-499
• /
• 2022

Machine learning such as deep learning have been widely used in recent years. Recently deep learning is performed in a trusted execution environment such as ARM TrustZone to improve security in edge devices and embedded devices with low computing resource. To mitigate this problem, we propose TPMP that efficiently uses the limited memory of TEE through DNN model partitioning. TPMP achieves high confidentiality of DNN by performing DNN models that could not be run with existing memory scheduling methods in TEE through optimized memory scheduling. TPMP required a similar amount of computational resources to previous methodologies.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.11
• /
• pp.377-386
• /
• 2020

Mobile cloud computing involves mobile or embedded devices as clients, and features small devices with constrained resource and low availability. Due to the fast expansion of smart phones and smart peripheral devices, researches on mobile cloud computing attract academia's interest more than ever. Computation offloading, or code offloading, enhances the performance of computation by migrating a part of computation of a mobile system to nearby cloud servers with more computational resources through wired or wireless networks. Code offloading is considered as one of the best approaches overcoming the limited resources of mobile systems. In this paper, we analyze the factors and the performance of code offloading, especially focusing on static program partitioning and data protection. We survey state-of-the-art researches on analyzed topics. We also describe directions for future research.

• Journal of agriculture & life science
• /
• v.46 no.6
• /
• pp.1-8
• /
• 2012

This study was carried out to determine the genic values of taste of Korean hot pepper (Capsicum annuum L.) in practical genetic resources of using to breed them. The two breeding materials of pepper, '#1803' ($P_1$) of prefer tastes and '#1532' ($P_2$) of ordinary taste, and their $F_1$, $F_2$ generations were used in this study. By using partitioning method it was possible to estimate, from the $F_2$ generation, the number of effective factor pairs differentiating the two parents. There were found to be differentiated by two effective factor pairs. In practical genetic resource of using to breed the Korean hot pepper, the heritance of pepper tastes showed that the $F_1$ was better than excellent parent by reason of over dominant, but some $F_2$ was better than both parent by transgressive segregation. As the result, the magnitude of genic effects of A-a gene in pepper tastes was 0.36, and B-b gene was 0.64. The tastes of Korean hot pepper showed a complex inheritance by interaction effect on the two non-allelic factors of 0.94 and secondary effect of 2.86 at the most.

• Journal of KIISE:Software and Applications
• /
• v.41 no.8
• /
• pp.569-579
• /
• 2014

In a tactical-edge environment, where multiple weapon resources are coordinated together via services, it is essential to make an efficient binding between an abstract service and a resource that are needed to execute composite services for accomplishing a given mission. However, the tactical network that is used in military operation has low bandwidth and a high rate of packet loss. Therefore, communication overhead between services must be minimized to execute composite services in a stable manner in the tactical network. In addition, a tactical-edge environment changes dynamically, and it affects the connectivity and bandwidth of the tactical network. To deal with these characteristics of the tactical network we propose two service-resource reallocation methods which minimize the communication overhead between service gateways and effectively manage neutralization of gateways during distributed service coordination. We compared the effectiveness of these two - methods in terms of total communication overhead between service gateways and resource-allocation similarity between the initial resource allocation and the reallocation result.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.7B
• /
• pp.685-696
• /
• 2004

Real-time multimedia applications that require large amount of bandwidth need resource reservation before starting service for providing the QoS(i.e., Quality of Service). To reserve resources in advance, each reservation request has to notify its expectation on the required amount of resources and service duration. Using this information, a resource manager can schedule advance reservations. However, most existing resource management systems are adopting straightforward call admission control process (i.e., only immediate reservation) by checking currently available resources without considering the service duration. Hence, the resource management system that supports advance reservation has to manage confliction caused by indefinite service duration of immediate reservation. Even though the separation of resource pool according to type of reservation can prevent the confliction, it causes low resource utilization. In this paper, we propose an effective resource management scheme that supports both immediate and advance reservations by sharing resources dynamically. Using network cost function, the proposed scheme determines and adaptively adjusts resource boundary according to the confliction rate by varying weight parameters. And also, we define user utility function to quantify user satisfaction based on how well the reserved resource is guaranteed during service time. Simulation results using NS-2 network simulator show that the proposed scheme can achieve better resource utilization with preferable QoS than other schemes like static resource partitioning.

• The Korean Journal of Ecology
• /
• v.18 no.4
• /
• pp.512-523
• /
• 1995

The central theme of Habital heterogeneity is to provide animals with habital complexity or structural diversity and to allow resource partitioning among individuals. In turn, the leads to population stability because prey can escape more easily with more hiding places causing less population fluctuation. Species diversity is characterized due to more potential niches both horizontally and verticall. Empirically, in homogeneous habitats population was less abundant, reproduction and survival were lower, spacing behavior, competition and dispersal were higher than in heterogeneous habitats. The results imply that diversity and conservation of species can be maintained through providing heterogeneous habitats.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.40 no.4
• /
• pp.30-44
• /
• 2003

In the microcell-or picocell-based system the frequent movements of mobiles bring about excessive traffics into the networks. A mobile location estimation mechanism can facilitate both efficient resource allocation and better QoS provisioning through handoff optimization. Existing location estimation schemes consider only LOS model and have poor performance in presence of multi-path and shadowing. In this paper we study novel schemes which can increase estimation accuracy by considering NLOS environment and other multiple decision parameters than the received signal strength.

• Journal of KIISE
• /
• v.42 no.6
• /
• pp.707-712
• /
• 2015

Mobile cloud computing is a technology that uses cloud services to overcome resource constrains of a mobile device, and it applies the computation offloading scheme to transfer a portion of a task which should be executed from a mobile device to the cloud. If the communication cost of the computation offloading is less than the computation cost of a mobile device, the mobile device commits a certain task to the cloud. The previous cost analysis models, which were used for separating functions running on a mobile device and functions transferring to the cloud, only considered the amount of data transfer and response time as the offloading cost. In this paper, we proposed a new task partitioning scheme that considers the frequency of function calls and data synchronization, during the cost estimation of the computation offloading. We also verified the energy efficiency of the proposed scheme by using experimental results.