• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.4 no.1
• /
• pp.68-75
• /
• 2011

The performance of peripheral devices is improving rapidly to meet the needs of users for multimedia data. Therefore, the peripheral interface with wide bandwidth and high transmission rate becomes necessary to handle large amounts of data in real time for multiple high-performance devices. PCI Express is a fast serial interface with the use of packets that are compatible with previous PCI and PCI-X. In this paper, we design and verify general peripheral interface using serial link. It includes two kinds of traffic class (TC) labels which are mapped to virtual channels (VC). The design adopts TC/VC mapping and the scheme of arbitration by priority. The design uses a packet which can be transmitted through up to four transmission lanes. The design of general peripheral interface is described in Verilog HDL and verified using ModelSim. For FPGA verification, Xilinx ISE and SPARTAN XC3S400 are used.We used Synopsys Design Compiler as a synthesis tool and the used library was MagnaChip 0.35um technology.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.7
• /
• pp.1603-1609
• /
• 2010

This Paper deals with a design, making a prtotype and test methods of Real-time towed Line Array Sensor Data processing board for fast data communication and long range data transmission with SFM(Serial FPDP Module) through Optic-fiber channel. Towed line array sensors are installed in Frigate and the each LAS A, B, C group data from LAS is packed a previously agreed protocol and transmitted to the signal processing unit. Considering the limited space of VME 6U size, LAS Data processing board is designed with MPC8265 PowerPC Controller of Freescale for main system control and Altera's CycloneIII FPGA for sensor data packing, self-test simulation data generation, S/W FIFO et cetera. LAS Data processing board have VxWorks, the RTOS(Real Time Operating System) that present many device drivers, peripheral control libraries on board for real-time data processing.

• Journal of KIISE:Computing Practices and Letters
• /
• v.10 no.1
• /
• pp.66-74
• /
• 2004

To get changing data or retrieved information in a web service environment, we need the CGI method in which we designate an executable file and receive the output of the executable file as a document. But the original CGI method has some performance pitfalls caused by the native condition that a process for the requested executable file should be created every time it is requested. Several approaches that make the process as a daemon wholly or partly have been made to solve those problems and they have exclusive strength and weakness. So, many developers are confused when they have to choose which of the CGI methods. We proposed an efficient CGI gateway design called SendFD that the main part of CGI gateway was separated as a daemon and taken over the communication channel by the other small part forked whenever it was requested. Then we testified the efficiency of SendFD styled CGI gateway in real web service environment and it showed about 3% performance improvement compared to the conventional socket pipe method.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.17 no.3
• /
• pp.327-335
• /
• 2014

In this paper, we propose a group distributed dynamic address assignment scheme suitable for tactical mobile ad hoc networks(MANET). Efficient address assignment is an important issue in the MANET because a node may frequently leave the current network and join another network owing to the mobility of the node. The conventional schemes do not consider the features of the tactical networks: existence of a leader node and network activity on a group basis. Thus, they may not be suitable for military operations. In our proposed scheme, called grouped units dynamic address assignment protocol(G-DAAP), a leader node maintains the address information for the members in the network and any of the nodes can exploit the information for the assignment or request of the IP address by a simple message exchange procedure. This leads to fast address assignment with small overheads. In addition, G-DAAP based on the modified IEEE 802.11e Enhanced Distributed Channel Access(EDCA) can assign addresses more quickly. We describe the delay performance of the G-DAAP and compare it with conventional schemes by numerical analysis and computer simulations. The results show that the G-DAAP significantly improves the delay performance as compared with the conventional schemes.

• Information Systems Review
• /
• v.8 no.3
• /
• pp.105-123
• /
• 2006

Traditional commerce has transformed into electronic commerce, then, mobile commerce again plays a critical role in diversifying different types of business. Although the fast changing of mobile commerce as an emerging channel of communication in business, not many studies address why consumers stay on mobile commerce and how they are influenced and influence other consumers. In this study, we investigate the impact of extrinsic and intrinsic motivation on customers' satisfaction and its results. We adopt the Technology Acceptance Model(TAM), value-satisfaction theories, and other motivation studies in developing theoretical framework. We then synthesize these theoretical approaches to conceptualize mobile customers' behavior. Specifically, we postulate cognitive beliefs that influence customers' satisfaction with mobile commerce, and, in turn, significantly contribute to the spread of positive word-of-mouth. The current research model and its results could provide a rigorous basis for understanding customers' behavior in wireless contexts.

• Proceedings of the Korea Contents Association Conference
• /
• 2005.05a
• /
• pp.507-513
• /
• 2005

This paper proposes a fast handoff scheme between PDSNs in 3GPPs network for a mobile node. It introduces a method by which handoff can be performed without reestablishing PPP connection that may occur in the process of performing handoff between PDSNs. The method for handoff between PDSNs which provide packet services to a mobile node, requires that the PDSNs should receive subscribers information about mobile nodes from their neighbor PDSNs forming a communication network. When the PDSN recognizes the mobile node moving into its coverage area, it can quickly establish a communication channel with the mobile node based on the already received subscriber information. As a result, handoff is performed without reestablishing PPP. Accordingly, handoff between PDSNs can be performed faster, removing time needed for establishing a PPP session with a terminal and for terminating a previously set up PPP session.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.2A
• /
• pp.165-173
• /
• 2010

WMN(Wireless Mesh Network) is an wireless backbone network technology that is an easily configurable network in the low cost compared to the wireless LAN(Local Area Network). Most of the previous researches have evaluated their algorithms by the simulations rather than by the implementation. There exist some implementation papers, however, they have the limitations of the flexibility on the link establishment and the link quality utilization. Consequently, the benefit of the WMN - configuration flexibility is degraded and the performance deterioration occurs in the multi-hop wireless environment. In this paper, we introduce a Linux-based link layer Wireless Mesh Routing System - WBMR. The design and implementation of WBMR provides the dynamic link establishment and the effective multi-channel usage. We have modified the ntroof the original WLAN operation for the dynamic link establishment, and the Linux bridge for the link layer routing. The result of performance evaluation verifies that our WBMR supports fast self-configuration and increases data transmission throughput compared to the other researches of the wireless multi-hop environment.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2002.11a
• /
• pp.368-372
• /
• 2002

Channel of distribution of contents of digitalised multimedia contents is changing in form passed directly to consumer from manufacturer through digital network of high speed. Together, diffusion of various digital equipment that do is changing Past multimedia consumption structure with the wonderful speed so that can consume easily digital contents. Specially, Multimedia Personal Computer and fast diffusion of a digital television are acting role such as catalyst that promote easy purchase and consumption of multimedia contents of good qualify. However, this system is no method that can sell digital contents and express right that original owner for the contents can insist nothing but consume. Because consumer can distribute to another person copying contents that buy because of and user can distribute producing ashes again according to necessary field by oneself as well as, can lose meaning for original contents. In this paper, Text editing system for XrML (extensible Rights Markup Language) that describe about integrity of message and entity authentication in addition to necessary rights, fees, condition etc. and this to do fetters XML (eXtensible Markup Language) for copyright protection for digital contents in treatise that see hereupon and use digital contents design and embody.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.9
• /
• pp.1957-1967
• /
• 1997

Adaptive channel equlization accomplished without resorting to a training sequence is known as blind equalization. In this paper, in order to reduce the speed of the convergence and the steady-state mean squared error simultaneously, we propose the multi-stage RD(radius-directed) algorithm derived from the combination of the constant modulus algorithm and the radius-directed algorithm. In the starting stage, multi-stage RD algorithm are identical to the constant modulus algorithm which guarantees the convergence of the equalizer. As the blind identical to the constant modulus algorithm which guarantees the convergence of the equalizer. As the blind equalizer converges, the number of the level of the quantizers is increased gradually, so that the proposed algorithm operate identical to the radius-directed algorithm which leads to the low error power after the covnergence. Therefore, the multi-stage RD algorithm obtains fast convergence rage and low steady stage mean square error.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.4A
• /
• pp.233-240
• /
• 2005

In this paper, we have shown the gain recovery time of erbium-doped fiber amplifier(EDFA) in WDM add/drop networks can be reduced to less than $3\mu$sec by employing disturbance observer technique with PID controller which is based on 3-level EBFA model. The $3\mu$sec transient time is the fastest one ever reported and it is about $1\%$ of $250\mu$sec previously obtained from disturbance observer technique with PI controller which was based on 2-level EDFA model and it is also less than $200\mu$sec of commercially available EDFAs for WDM networks. We were able to obtain this result by analyzing various system characteristics as we change the parameters of disturbance observer and PID controller. If this result is applied to a commercial product, the product will do an important role in the future dynamic WDM networks.