• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10e
• /
• pp.628-630
• /
• 2002

기가급 이상의 고속 인터넷 접속이 필수적인 고성능 리눅스 서버 시스템에 사용되는 네트워크 프로토콜의 성능향상을 위하여, 기존의 소프트웨어 성능향상 기술과 고속 패킷 처리 전용의 네트워크 프로세서를 활용하여 고속 네트워크 프로토콜을 설계 및 구현하였다. 네트워크 정합 장치로는 인텔의 IXP1200 네트워크 프로세서가 탑재된 네트워크 카드를 사용하였고, 리눅스 커널의 TCP／UDP／IP 네트워크 프로토콜 스택 중 계층 3 IP 프로토콜 이하를 네트워크 프로세서가 처리하게 함으로써 커널의 오버헤드를 줄이고, 고속의 프로토콜 처리가 가능하게 되었다.

• Korean Journal of Remote Sensing
• /
• v.17 no.3
• /
• pp.243-251
• /
• 2001

GPS is one of today's most widely used surveying techniques. But, users can't acquire an enough accuracy in applications of the navigation or geodesy by the GPS positioning technique because of the effects of the ionosphere and troposphere. The solution of these restrictions in the DGPS technique that is to eliminate the common errors and can achieve a high accuracy. Although of sufficient density for good DGPS, accuracy of positioning is just not dense enough to provide complete coverage for real-time positioning, because distances between base and rover is short. In this paper, we suggest Realtime Long-Distance Transmission Method of DGPS Error Correction Signal, which consist of TCP, UDP and IP, which allows a user to increase the distance at which the rover receiver is located from the base, due to radio modem.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.3
• /
• pp.493-497
• /
• 2005

This paper is about the research of the User Agent implementation under wireless embedded environment, using SIP which is one of protocol components construct the VoIP system. The User Agent is made of the User Agent configuration block, the device thread block to control devices and the SIP stack block to process SIP messages. The device thread consists of the RTP thread and the sound lard device processing block. Futhermore, the SIP stack consist of the worker thread to process proxy events, the SIP transceiver and SIP thread to transfer and receive SIP messages. The H/W platform is a board included the Intel's XScale PXA255 processor, flash memory, SDRAM, Audio CODEC module and wireless LAN threough PCMCIA socket, furthermore a microphone and headphone is used by the audio 1/0. The system has embedded linux kernel 2.4.19. For embedded environment, the function of User Agent and SIP method is diminished. Finally, the resource of system could be reduced about $12.9\%$, compared to overall system resource, by minimizing peripherals control and excepting TCP.

• Proceedings of the IEEK Conference
• /
• 2003.11b
• /
• pp.53-56
• /
• 2003

Ethernet is one of the most successful LAN technologies. Now gigabit ethernet is available in real network and some network interface cards(NIC) supports TCP segment offloading (TSO), IP checksum offloading(ICO), Jumbo frame and interrupt moderation. If we use this features appropriately, we obtain high throughput with low CPU utilization. This paper represents the network performance by varying above features.

• Journal of Advanced Navigation Technology
• /
• v.24 no.5
• /
• pp.438-443
• /
• 2020

To transfer files between nodes on network based on Ethernet, file transfer protocol (FTP) on TCP/IP and trivial file transfer protocol (TFTP) on UDP/IP are mostly used. Due to the lack of resources (processor, memory and so on) in the embedded system where we generally use for simple works with small firmware like ones; many of the systems implement only UDP/IP for their network stacks. Thus, TFTP is greatly to be preferred. For examples, environmental sensor devices for sensor networks, Boot Loader for general embedded device and preboot execution environment (PXE) boot for PC provide the TFTP. The logic of TFTP is simple for file transmission but, there is Stop-And-Wait problem during the process which occurs long blocking time. In this paper, we propose an algorithm which called QuTFTP(Quick UDP Trivial File Transfer Protocol) to reduce the length of the blocking time and to be compatible with the legacy TFTP.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.6B
• /
• pp.575-583
• /
• 2004

In this paper, we extend the multiple time scale control framework to the window-based congestion control, in particular, such as the TCP. This is performed by interfacing the TCP with a large time scale control module which adjusts the aggressiveness of the bandwidth consumption behavior exhibited by the TCP as a function of large time scale Self-Similar network state. i.e., conformation that exceeds the horizon of the feedback loop as determined by the RTT. How to effectively utilize such an information-due to its probabilistic nature, dispersion over the multiple time scales, and affection on the top of the existing window-based congestion controls-is a non-trivial problem. The evaluation performance of the multiple time scale TCP is facilitated by a simulation of the bench-mark environment which is based on the physical modeling of a self-similar traffic. We explicate our methodology for discerning and evaluating the impact of changes in transport protocols in the protocol stack under the self-similar traffic conditions. We discuss issues arising in the comparative performance evaluation under heavy-tailed workloads.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.3
• /
• pp.549-557
• /
• 2017

ICMP(Internet Control Message Protocol) is a main protocol in TCP/IP protocol stack. ICMP compensates the disadvantages of the IP that does not support error reporting. If any transmission problem occurred, a router or receiving host sends ICMP message containing the error cause to sending host. However, in this process, an attacker sends a fake ICMP messages to the host so that the communication can be terminated abnormally. An attacker host can paralyzes system of victim host by sending a large number of messages to the victim host at a high rate of speed. To solve this problem, we have designed and implemented outbound traffic controller that prevents various ICMP attacks. By preventing the transmission of attack messages in different ways according to each case, various network attacks can be prevented. In addition, unnecessary network traffic can be filtered before transmitted.

• Journal of KIISE:Computing Practices and Letters
• /
• v.10 no.6
• /
• pp.460-471
• /
• 2004

Since the RTP is suitable for real-time data transmission in multimedia services like VoD, AoD, and VoIP, it has been adopted as a real-time transport protocol by RTSP, H.323, and SIP. Even though the RTP protocol stack for embedded systems has been in great need for efficient support of multimedia services, such a stack has not been developed yet. In this paper, we explain embeddedRTP which supports the RTP protocol stack at the kernel level so that it is suitable for embedded systems. Since embeddedRTP is designed to reside in the UBP module, existing applications which rely ell TCP/IP services can proceed the same as before, while applications which rely on the RTP protocol stack can request HTP services through embeddedRTp API. EmbeddedRTP stores transmitted RTP packets into per session packet buffer, using the packet's port number and multimedia session information. Communications between applications and embeddedRTP is performed through system calls and signal mechanisms. Additionally, embeddedRTP API makes it possible to develop applications more conveniently. Our performance test shows that packet-processing speed of embeddedRTP is about 7.5 times faster than that oi VCL RTP for multimedia streaming services on PDA in spite that its object code size is reduced about by 58% with respect to UCL RTP's.

• Proceedings of the Korean Information Science Society Conference
• /
• 2008.06b
• /
• pp.355-360
• /
• 2008

대부분의 임베디드 시스템들은 유선 및 무선 네트워크에 연결되어 있으며 이들이 생성하는 데이터는 개인, 기관, 또는 국가에 상당히 중요한 정보일 수 있다. 따라서 통신 시 보안에 대한 요구는 향후 더욱 크게 증가할 것으로 전망된다. 임베디드 시스템에서의 또 다른 중요한 요구사항 중의 하나는 저전력 특성이다. 따라서 중요 데이터의 안전한 전송을 위한 암호화 및 복호화 그리고 네트워크 프로토콜 스택의 수행에 따른 전력 사용량 정보는 차세대 저전력 임베디드 시스템을 설계 및 개발하기 위해서 필수적으로 요구될 것이다. 하지만 기존 연구들은 단편적인 전력 사용량만을 분석하고 있다. 본 논문은 보안 데이터를 위한 암호화, 복호화, 그리고 데이터 송수신을 위한 네트워크 프로토콜 스택의 수행에 걸리는 전력 사용량과 오버헤드를 함께 측정하고 분석한다. 측정 결과 3DES 기반의 암호화 및 복호화 작업 동안의 전력 사용량이 TCP/IP 통신을 위한 전력 사용량 보다 데이터 크기가 증가함에 따라 더욱 크게 측정되었다. 해당 작업의 수행 오버헤드 역시 유사한 성향을 보였다. 그리고 프로세서 코어의 전력 사용량은 시스템 전체 사용량의 $10{\sim}22%$ 정도를 차지하는 것으로 측정되었다. 이러한 연구 결과는 임베디드 시스템에서 저전력 보안 데이터 전송을 위해 고려할 사항들을 정량적으로 제시하여 해당 연구 분야에 기여할 수 있다.

• Journal of KIISE:Information Networking
• /
• v.34 no.1
• /
• pp.16-26
• /
• 2007

It is well known that, in the near future, the lifetime of the IPv4 address space will be limited and available 32-bit IP network addresses will not be left my more. In order to solve such IPv4 address space problem in an effective way, the transition to the new version using IPv6 architecture is inevitably required. This paper presents the design and implementation of IPv4/IPv6 dual stack at the GSM Phone based on Linux Kernel 2.4 IPv6 Protocol Stack. It designs appropriately in GSM Phone environment and it is tested by a network of Linux IPv4/IPv6 dual stack on PPP. The test was processed with a test scenario and it was found that the results were successful.