Praphan Pavarangkoon

A quality measurement tool for high-speed data transfer in long fat networks

Due to rapid increase of network bandwidth, applications and systems working in long fat network (LFN) play more important roles. For effective development of them, precise measurements of network conditions are significant. Transmission Control Protocol (TCP) is the most commonly used protocol, but is essentially unable to achieve high throughput in LFNs with packet losses. For this reason, it is hard for conventional network measurement tools to show the maximum or available bandwidth in LFN, especially in high packet loss environments. To overcome this issue, we introduce a novel data transfer protocol on TCP/IP transport layer, namely high-performance and flexible protocol (HpFP). For high-precision pace control and retransmission control, the HpFP intermittently monitors network conditions such as packet loss and latency. We develop an application via the HpFP, named hperf, which measures end-to-end throughput as well as status of packet loss and latency in LFNs. We carry out experiments to examine the abilities of the hperf in high-throughput data transfer and measurement of network qualities in terms of packet loss and latency. The hperf achieves almost wire-rate throughput, 10 Gbps, on the international link between Japan and the USA with even 0.5% packet loss ratio (PLR). The measurements of packet loss and latency show good correspondence with the conventional methods via iperf and ping. These results are verified in our laboratory experiments on 10 Gbps link using a network simulator as well. We conclude that the HpFP has significant potential for a variety of network applications and the hperf is a good network quality measurement tool in LFNs, compared to the conventional TCPs.

Improvement of real-time transfer of phased array weather radar data on long-distance networks

With the tremendous development of remote sensing technologies, large-scale data sets are collected from environmental and meteorological sensors. Edge computing for large-scale data sets is hard to be conducted in real time on the sensors, thus data transfer to cloud system and data processing play an important role. One of the barriers in such system is high-speed data transfer on long-distance networks. Even low packet loss condition leads to decrease in throughput by conventional transmission control protocol (TCP)-based tools. In this paper, we overcome this issue by proposing a high-performance data transfer tool to improve real-time data processing of phased array weather radar on long-distance networks. The proposed tool, called high-performance copy (hcp), is designed based on an open source tool for file copy operations and implemented using our socket library specialized for long-distance networks. We evaluate our tool with simulations and real data transfer scenarios. The results show that the hcp enables concurrent data transfer from radar sites to cloud system (data processing site) in real time. The throughput of data transfer from a radar site to a data processing site increases from 40 Mbps using the conventional TCP-based tool to 489.2 Mbps using the hcp.

A web-based real-time and full-resolution data visualization for Himawari-8 satellite sensed images

It has been almost four decades since the first launch of geostationary meteorological satellite by Japan Meteorological Agency (JMA). The specifications of the geostationary meteorological satellites have shown tremendous progresses along with the generations, which are now entering their third generation. The third-generation geostationary meteorological satellites not only yield basic data for weather monitoring, but also globally observe the Earth’s environment. The development of multi-band imagers with improved spatial resolution onboard the third-generation geostationary meteorological satellites brings us meteorological data in larger size than those of the second-generation ones. Thus, new techniques for domestic and world-wide dissemination of the observational big data are needed. In this paper, we develop a web-based data visualization for Himawari-8 satellite sensed images in real time and with full resolution. This data visualization is supported by the ecosystems, which uses a tiled pyramid representation and parallel processing technique for terrain on an academic cloud system. We evaluate the performance of our techniques for domestic and international users on laboratory experiments. The results show that our data visualization is suitable for practical use on a temporal preview of observation image data for the domestic users.

Real-time 3D visualization of phased array weather radar data via concurrent processing in Science Cloud

With the tremendous development of remote sensing technologies, a large amount of observation data are generated from sensors. Since each sensor generates data periodically, e.g., every minute, a concurrent data processing using a cloud system plays an important role in the modern design process. This paper focuses on concurrent data processing techniques for an X-band phased array weather radar (PAWR) using high-speed network, parallel data processing system, and large-scale storage system. The PAWR at National Institute of Information and Communications Technology (NICT), Japan rotates in 30 sec to capture a 3D structure of rainfalls within 60 km in radius and 15 km in altitude. In this paper, we develop a real-time 3D visualization system of the observation data of the PAWR. Our visualization is carried out from 54 sec to 69 sec (depending on the weather conditions) after every observation period, which is in the same time scale with other conventional 2D visualization of X-band weather radars. In addition, we discuss a combination of cloud ecosystems for the concurrent processing at low cost. The methodology is considered as a pioneering case study to develop of a variety of real-time data processing systems of big data via remote sensing.

An application of novel communications protocol to high throughput satellites

For network communications using modern high throughput satellite (HTS) on geostationary orbits, network throughput of transmission control protocol (TCP), one of the most popular protocols, is limited due to the packet loss on the satellite link. The packet loss is mainly caused by the attenuation of signals in severe weather conditions like heavy rain. It is high time to develop novel network communication techniques on the transport layer in TCP/IP designed for the systems and applications in broadband communications. In this paper, we introduce a high-speed data transfer protocol, named high-performance and flexible protocol (HpFP), to achieve high throughput for the HTS even with packet loss. The HpFP, in comparison with TCP-Hybla and UDP-based data transfer (UDT) protocols, is evaluated on a laboratory experiment simulating a geostationary orbit satellite link of 10 Gbps. It is clarified that the HpFP outperforms both the TCP-Hybla and the UDT showing high throughputs (close to 10 Gbps) when the packet loss ratio (PLR) is 1%, and remains more than 1 Gbps under even 10% PLR condition. Moreover, in case of no packet loss, the HpFP exhibits a quick start-up time (6 sec) at the initial phase to achieve 10 Gbps, while the TCP-Hybla and the UDT take 9 sec and 16 sec to their maximum throughputs, respectively.

A high-speed data transfer protocol for geostationary orbit satellites

In communication systems using geostationary orbit satellites, throughput of transmission control protocol (TCP) is limited due to the impact of latency on network and packet loss caused by signal attenuation in severe weather conditions like heavy rain. It is high time to develop network techniques and applications in broadband communications over the gigabit satellite and the high throughput satellite (HTS). In this paper, we introduce a high-speed data transfer protocol, named high-performance and flexible protocol (HpFP), to achieve high throughput over a geostationary satellite link even in severe weather conditions. The HpFP is firstly evaluated on a laboratory experiment simulating a geostationary orbit satellite link. It is clarified that the HpFP shows high throughputs even when the packet loss ratio (PLR) is 0.01%. We next carry out a field experiment using the Wideband InterNetworking engineering test and Demonstration Satellite (WINDS). The performance of the HpFP over single, dual, and multiple connections are evaluated. The result shows that the aggregate throughput of dual connections of HpFP almost reaches to the maximum bandwidth, and the time to the maximum bandwidth is within 3 sec which is over 20 times faster than that by the TCP. For multiple connections, the HpFP shares the bandwidth equally among all 50 connections.

Design and Development of Real-Time Video Transmission System Using Visual IoT Device

Visual Internet of Things (IoT) is a class of IoT that collects rich visual data over the Internet. In general, the visual IoT device is equipped with video transmission equipment such as a mobile camera. Both advanced video transmission techniques and information extraction from images by image recognition techniques are key techniques for the visual IoT. However, since the video data size is larger than the sensor data size in general, one of the issues of visual IoT is high-performance video transmission in networks in which the bandwidths are limited. In this paper, we design a real-time video transmission system using visual IoT device. Our system is based on a novel protocol, named high-performance video transmission (HpVT), for field monitoring via 4G LTE mobile networks. Our implementation of the system is based on Raspberry Pi boards, which are single-board computers with ARM processor. We evaluate the performance of our system in real fields to conclude that we can achieve full high-definition (full HD) resolution video transmission with as high frame rate as 30 fps even from a vehicle moving on a highway.

Correction to: A web-based real-time and full-resolution data visualization for Himawari-8 satellite sensed images

The original version of this article, unfortunately, contained errors. In Tables 3 and 6, there were entries that were not properly aligned on its designated rows.

Multiple streams of UDT and HpFP protocols for high-bandwidth remote storage system in long fat network

Large-scale distributed file system is an important technology for big data sciences. Several works so far have succeeded in high-bandwidth data I/O on multi-point to multi-point remote storage systems. However, for easy use by many scientific researchers, high-bandwidth remote storage tools working on a single client computer are preferable. As for high-bandwidth multi-point to single-point remote storage system, there is a study to introduce a high-bandwidth virtual remote storage (HbVRS) tool using a distributed file system (Gfarm) and a UDP-based data transfer protocol (UDT) in long fat network (LFN). In this paper, we introduce a novel transport protocol, named high-performance and flexible protocol (HpFP), to show high throughputs for the HbVRS in LFNs with packet loss. We examine basic performances of multiple streams of the HpFP and the UDT on a laboratory experiment simulating an international LFN. In the presence of packet losses (0.5%), the HpFP shows more than 9.3 Gbps network throughput at any parallelization numbers from 1 to 12, while the maximum throughput by the UDT is limited in 1.4 Gbps with 12 parallelization.

Bandwidth utilization enhancement using high-performance and flexible protocol for INTELSAT satellite network

Achieving the quality of service (QoS) is an important requirement in a communication network. Satellite communication is posing many challenges due to the limitation of transmission control protocol (TCP) over networks with high latency. To overcome these issues, the wide area network (WAN) optimization provides the data transfer on such long-distance networks. However, this optimization is not able to utilize the available bandwidth of provided network efficiently since it performs fixed bandwidth allocation. This paper proposes a technique to enhance the available bandwidth utilization for International Telecommunications Satellite Organization (ITSO, or INTELSAT) network. This technique adopts a high-speed data transfer protocol, named high-performance and flexible protocol (HpFP), to transfer data between the satellite and the ground station. The HpFP is a connection-oriented protocol to work on the top of user datagram protocol (UDP) and provides us with a stream-type of reliable data transfer even under high packet loss rate. One of the ingenious attempts in the HpFP is to set an internal target throughput for pace control of sending packets. Since this parameter setting is time-dependent, the target throughput is calculated based on network conditions monitored by the HpFP. The HpFP detects the unused bandwidth in the satellite bandwidth resources at every moment, then dynamically allocates HpFP data transfers. The results of laboratory experiments show how effectively the HpFP utilizes the available network bandwidth in the condition of the WAN optimization control on INTELSAT satellite network.