• The KIPS Transactions:PartC
• /
• v.12C no.4 s.100
• /
• pp.571-580
• /
• 2005

In fast long-distance networks, TCP's congestion control algorithm has the problem of utilizing bandwidth effectively. Several window-based congestion control protocols for high-speed and large delay networks have been proposed to solve this problem. These protocols deliberate mainly three properties : scalability, TCP-friendliness, and RTT-fairness. These protocols, however, cannot satisfy above three properties at the same time because of the trade-off among them This paper presents a new window-based congestion control algorithm, called EM (Exponential Increase/ Multiplicative Decrease), that simultaneously supports all four properties including fast convergence, which is another important constraint for fast long-distance networks; it can support scalability by increasing congestion window exponentially proportional to the time elapsed since a packet loss; it can support RTT-fairness and TCP-friendliness by considering RTT in its response function; it can support last fair-share convergence by increasing congestion window inversely proportional to the congestion window just before packet loss. We evaluate the performance of EIMD and other algorithms by extensive computer simulations.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.1
• /
• pp.78-101
• /
• 2017

High-quality industrial control is critical to ensuring production quality, reducing production costs, improving management levels and stabilizing equipment and long-term operations. WiFi-based Long Distance (WiLD) networks have been used as remote industrial control networks. Real-time performance is essential to industrial control. However, the original mechanism of WiLD networks does not minimize end-to-end delay and restricts improvement of real-time performance. In this paper, we propose two algorithms to obtain the transmitting/receiving phase cycle length for each node such that real time constraints can be satisfied and phase switching overhead can be minimized. The first algorithm is based on the branch and bound method, which identifies an optimal solution. The second is a fast heuristic algorithm. The experimental results show that the execution time of the algorithm based on branch and bound is less than that of the heuristic algorithm when the network is complex and that the performance of the heuristic algorithm is close to the optimal solution.

• Journal of Communications and Networks
• /
• v.12 no.3
• /
• pp.275-284
• /
• 2010

The current multihome-aware protocols (like stream control transmission protocol (SCTP) or parallel TCP for concurrent multipath data transfer (CMT) are not designed for high-capacity and large-latency networks; they often have performance problems transferring large data files over shared long-distance wide area networks. It has been shown that SCTP-CMT is more sensitive to receive buffer (rbuf) constraints, and this rbuf-blocking problem causes considerable throughput loss when multiple paths are used simultaneously. In this research paper, we demonstrate the weakness of SCTP-CMT rbuf constraints, and we then identify that rbuf-blocking problem in SCTP multihoming is mostly due to its loss-based nature for detecting network congestion. We present a simulation-based performance comparison of FAST TCP versus SCTP in high-speed networks for solving a number of throughput issues. This work proposes an end-to-end transport layer protocol (i.e., FAST TCP multihoming as a reliable, delaybased, multihome-aware, and selective ACK-based transport protocol), which can transfer data between a multihomed source and destination hosts through multiple paths simultaneously. Through extensive ns-2 simulations, we show that FAST TCP multihoming achieves the desired goals under a variety of network conditions. The experimental results and survey presented in this research also provide an insight on design decisions for the future high-speed multihomed transport layer protocols.

• Journal of Communications and Networks
• /
• v.6 no.1
• /
• pp.1-8
• /
• 2004

In this paper, we investigate the properties and performance of space-time bit-interleaved coded modulation (STBICM) systems in fast Rayleigh fading channels. We first show that ST-BICM with QPSK signaling in fast fading channels possesses the uniform distance property, which makes performance analysis tractable. We also derive the probability distribution of the squared Euclidean distance between space-time symbols assuming uniform bit-interleaving. Based on the distribution, we show that the diversity order for each codeword pair becomes maximized as the frame length becomes sufficiently long. This maximum diversity order property implies that the bit-interleaver transforms an ST-BICM system over transmit diversity channels into an equivalent coded BPSK system over independent fading channels. We analyze the performance of ST-BICM in fast fading channels by deriving an FER upper bound. The derived bound turns out very accurate, requiring only the distance spectrum of the binary channel codes of ST-BICM. Numerical results demonstrate that the bound is tight enough to render an accurate estimate of performance of ST-BICM systems.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2005.05a
• /
• pp.1271-1274
• /
• 2005

Fast long-distance network 에서 기존 TCP 의 혼잡 제어 (congestion control) 알고리즘은 대역폭을 효과적 사용하지 못하는 문제점을 가지고 있다. 대역폭을 효과적으로 사용하기 위해서 TCP 혼잡 제어를 수정한 다양한 프로토콜들이 제안되었다. 이러한 프로토콜들은 디자인 시 주로 bandwidth scalability, TCP friendliness, 그리고 RTT fairness 와 같은 세 가지의 특성을 고려하고 있다. 하지만 제안된 프로토콜들은 어떤 것도 trade-off 관계로 있는 이 세 가지 특성을 동시에 만족시키지 못한다. 본 논문에서는 혼잡 제어 알고리즘의 증가 규칙 (increase rule)에 RTT 를 직접 반영함으로써 위 세가지 요구사항을 동시에 만족시키는 EIMD (Exponential Increase/ Multiplicative Decrease)라고 하는 새로운 TCP 혼잡 제어 알고리즘을 제안한다. EIMD 는 패킷 손실이 없는 한, 지수적으로 윈도우를 증가시켜 효과적으로 대역폭을 사용하면서도, 패킷손실 직전의 윈도우 크기, $W_{max}$ 에 반비례하게 윈도우를 증가시킴으로써 fair share 에 빠르게 수렴할 수 있다는 특성을 갖는다. 모의실험을 통해 제안된 프로토콜이 fast long-distance network 에서 위 4 가지 특성들을 모두 만족하는지 검증한다

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.43 no.11 s.353
• /
• pp.78-84
• /
• 2006

This paper proposes a method to endow the mobility to the Zigbee protocol which is a short distance wireless communication technology among the standards of wireless sensor networks. The Zigbee has a low price, long life battery, simple structure and connectivity. So it is a suitable standard for the short distance wireless communication of $10m{\sim}70m$. As the Zigbee does not support the mobility, additional factors should be studied to deal adeptly with the diverse change in the real field. By solving the problem that the previous connection and binding table of the Zigbee end device should be reset due to the Zigbee end device address variation when the Zigbee end device moves from the present router to another router, this paper presents the method to communicate without data loss or disconnection.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.4
• /
• pp.467-478
• /
• 2016

In this paper, we try to find the optimal locations of NeNB(Nomadic evolved NodeB)s for maximizing the overall throughput of the PS-LTE networks. Since finding optimal locations of all NeNBs in a given area is NP-hard(Non-deterministic Polynomial time-hard) problem, we proposed a PSO-based heuristic approach. In order to evaluate the performance, we conducted two experiments. We compared performance with other schemes such as Exhaustive Search, Random Walk Search, and locating neighboring NeNBs with the same NeNB-to-NeNB distance. The proposed method showed the similar results to the exhaustive search method in terms of locating optimal position and user's data throughput. The proposed method, however, has the fast and consistent convergence time.

• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.1-25
• /
• 2020

In this paper, we suggest an application system architecture which provides accurate, fast and efficient automatic gasometer reading function. The system captures gasometer image using mobile device camera, transmits the image to a cloud server on top of private LTE network, and analyzes the image to extract character information of device ID and gas usage amount by selective optical character recognition based on deep learning technology. In general, there are many types of character in an image and optical character recognition technology extracts all character information in an image. But some applications need to ignore non-of-interest types of character and only have to focus on some specific types of characters. For an example of the application, automatic gasometer reading system only need to extract device ID and gas usage amount character information from gasometer images to send bill to users. Non-of-interest character strings, such as device type, manufacturer, manufacturing date, specification and etc., are not valuable information to the application. Thus, the application have to analyze point of interest region and specific types of characters to extract valuable information only. We adopted CNN (Convolutional Neural Network) based object detection and CRNN (Convolutional Recurrent Neural Network) technology for selective optical character recognition which only analyze point of interest region for selective character information extraction. We build up 3 neural networks for the application system. The first is a convolutional neural network which detects point of interest region of gas usage amount and device ID information character strings, the second is another convolutional neural network which transforms spatial information of point of interest region to spatial sequential feature vectors, and the third is bi-directional long short term memory network which converts spatial sequential information to character strings using time-series analysis mapping from feature vectors to character strings. In this research, point of interest character strings are device ID and gas usage amount. Device ID consists of 12 arabic character strings and gas usage amount consists of 4 ~ 5 arabic character strings. All system components are implemented in Amazon Web Service Cloud with Intel Zeon E5-2686 v4 CPU and NVidia TESLA V100 GPU. The system architecture adopts master-lave processing structure for efficient and fast parallel processing coping with about 700,000 requests per day. Mobile device captures gasometer image and transmits to master process in AWS cloud. Master process runs on Intel Zeon CPU and pushes reading request from mobile device to an input queue with FIFO (First In First Out) structure. Slave process consists of 3 types of deep neural networks which conduct character recognition process and runs on NVidia GPU module. Slave process is always polling the input queue to get recognition request. If there are some requests from master process in the input queue, slave process converts the image in the input queue to device ID character string, gas usage amount character string and position information of the strings, returns the information to output queue, and switch to idle mode to poll the input queue. Master process gets final information form the output queue and delivers the information to the mobile device. We used total 27,120 gasometer images for training, validation and testing of 3 types of deep neural network. 22,985 images were used for training and validation, 4,135 images were used for testing. We randomly splitted 22,985 images with 8:2 ratio for training and validation respectively for each training epoch. 4,135 test image were categorized into 5 types (Normal, noise, reflex, scale and slant). Normal data is clean image data, noise means image with noise signal, relfex means image with light reflection in gasometer region, scale means images with small object size due to long-distance capturing and slant means images which is not horizontally flat. Final character string recognition accuracies for device ID and gas usage amount of normal data are 0.960 and 0.864 respectively.