• Journal of Korea Society of Digital Industry and Information Management
• /
• v.10 no.3
• /
• pp.35-40
• /
• 2014

Many researches have been conducted to achieve improvement in distributed system that connects multiple computer systems via communication lines. Among others, the load balancing and file migration are considered to have significant impact on the performance of distributed system. The dynamic file migration algorithm common in distributed processing system involved complex calculations of decision function necessary for file migration and required migration of control messages for the performance of decision function. However, the performance of this decision function puts significant computational strain on computer. As one single network is shared by all computers, more computers connected to network means migration of more control messages from file migration, causing the network to trigger bottleneck in distributed processing system. Therefore, it has become imperative to carry out the research that aims to reduce the number of control messages that will be migrated. In this study, the learning automata was used for file migration which would requires only the file reference-related information to determine whether file migration has been made or determine the time and site of file migration, depending on the file conditions, thus reflecting the status of current system well and eliminating the message transfer and additional calculation overhead for file migration. Moreover, mathematical model for file migration was described in order to verify the proposed model. The results from mathematical model and simulation model suggest that the proposed model is well-suited to the distributed system.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.18 no.33
• /
• pp.1-10
• /
• 1995

The allocation of files is essential to the efficiency and effectiveness of a distributed system that must meet geographically dispersed data processing demands. In this paper, we address an optimization model that generates optimal file migration policies in distributed database systems. The proposed model is a more generalized model that includes system's capacity constraints - computing sites' storage capacity and communication networks' capacity - which have not taken into consideration in previous researches. Using this model, we can establish initial file allocation, file reallocation and file migration polices that minimize a system operating cost under system's capacity constraints at an initial system design or reorganization point The proposed model not only can be adopted by small-sized systems but also provides a foundation for effective and simple heuristics for adaptive file migration in large systems.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.13 no.4
• /
• pp.153-162
• /
• 2017

Advances in communication technologies and the decreasing cost of computers have made distributed computer systems an attractive alternative for satisfying the information needs of large organizations. This paper presents a distributed algorithm for performance improvement through load balancing and file migration in distributed systems. We employed a sender initiated strategy for task migration and used learning automata with several internal states for file migration. A task can be migrated according to the load information of a computer. A file is migrated to the destination processor when it is in the right boundary state. We also described an analytical model for load balancing with file migration to verify the proposed algorithm. Analytical and simulation results show that our algorithm is very well-suited for distributed system environments.

• KIPS Transactions on Computer and Communication Systems
• /
• v.7 no.11
• /
• pp.275-280
• /
• 2018

The choice of a suitable distributed file system is required for loading large data and high-speed processing through subsequent applications in a cloud environment. In this paper, we propose a write performance improvement method based on GlusterFS and evaluate the performance of MapRFS, CephFS and GlusterFS among existing distributed file systems in cloud environment. The write performance improvement method proposed in this paper enhances the response time by changing the synchronization level used by the synchronous replication method from disk to memory. Experimental results show that the distributed file system to which the proposed method is applied is superior to other distributed file systems in the case of sequential write, random write and random read.

• Journal of KIISE:Computer Systems and Theory
• /
• v.36 no.4
• /
• pp.259-273
• /
• 2009

Most of large-scale distributed file systems decouple a metadata operation from read and write operations for a file. In the distributed file systems, a certain server named a metadata server (MDS) maintains metadata information in file system such as access information for a file, the position of a file in the repository, the namespace of the file system, and so on. But, the existing systems used restrictive metadata management schemes, because most of the distributed file systems designed to focus on the distributed management and the input/output performance of data rather than the metadata. Therefore, in the existing systems, the metadata throughput and expandability of the metadata server are limited. In this paper, we propose a new non-shared metadata management scheme in order to provide the high metadata throughput and scalability for a cluster of MDSs. First, we derive a dictionary partitioning scheme as a new metadata distribution technique. Then, we present a load balancing technique based on the distribution technique. It is shown through various experiments that our scheme outperforms existing metadata management schemes in terms of scalability and load balancing.

• Asia pacific journal of information systems
• /
• v.2 no.2
• /
• pp.71-84
• /
• 1992

Analysis for the file with distributed free space such as a VSAM has been done. Birth and death process has been adopted to describe the status changes of control intervals in a VSAM file. File utilization is calculated as a function of the number of control intervals which contain different number of records in them. Effect of the control interval sizes and the loading factors are analyzed over the time horizon.

• Journal of the Korea Society for Simulation
• /
• v.17 no.4
• /
• pp.229-239
• /
• 2008

A Distributed File System (DFS) provides a mechanism in which a file can be stored across several physical computer nodes ensuring replication transparency and failure transparency. Applications that process large volumes of data (such as, search engines, grid computing applications, data mining applications, etc.) require a backend infrastructure for storing data. And the distributed file system is the central component for such storing data infrastructure. There have been many projects focused on network computing that have designed and implemented distributed file systems with a variety of architectures and functionalities. In this paper, we describe a complete distributed file system which can be used in large-scale search-oriented systems.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.19 no.40
• /
• pp.169-178
• /
• 1996

Geographically dispersed computing system is made of computers interconnected by a telecommunications network. To make the system operated efficiently, system designer must determine the allocation of data files to each node. In designing such distributed computing system, the most important issue is the determination of the numbers and the locations where database files are allocated. This is commonly referred to as the file allocation problem (FAP)[3]. The proposed model is a 0/l integer programming problem minimizing the sum of file storage costs and communication(query and update) costs. File allocation problem belongs to the class of NP-Complete problems. Because of the complexity, it is hard to solve. So, this paper presents an efficient heuristic algorithm to solve the file allocation problem using Tabu Search Technique. By comparing the optimal solutions with the heuristic solutions, it is believed that the proposed heuristic algorithm gives good solutions. Through the experimentation of various starting points and tabu restrictions, this paper presents fast and efficient method to solve the file allocation problem in the distributed computing system.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.24 no.64
• /
• pp.1-20
• /
• 2001

In distributed databases, file replication and workload allocation are important design issues. This paper solves these two issues simultaneously, The primary objective is to minimize the system response time that consists of local processing and communication overhead on a local area network. Workload (query transactions) is assigned among any sites in proportion to the remaining file request service rate of the each server The problem is presented in the form of a nonlinear integer programming model. The problem is proved to be NP-complete and thus an efficient heuristic is developed by employing its special structure. To illustrate its effectiveness, it is shown that the proposed heuristic is based on the heuristic of a non-redundant allocation that was provided to be effective. The model and heuristics are likely to provide more effective distributed database designs.

• KIPS Transactions on Computer and Communication Systems
• /
• v.3 no.12
• /
• pp.433-442
• /
• 2014

In this paper, we introduce the client side cache of distributed file system for enhancing read performance by eliminating the network latency and decreasing the back-end storage burden. This performance enhancement can expand the fields of distributed file system to not only cloud storage service but also high performance storage service. This paper shows that the distributed file system with client side SSD cache can satisfy the requirements of VDI(Virtual Desktop Infrastructure) storage. The experimental results show that full-clone is more than 2 times faster and boot time is more than 3 times faster than NFS.