• Journal of Intelligence and Information Systems
• /
• v.26 no.4
• /
• pp.127-148
• /
• 2020

The data center is a physical environment facility for accommodating computer systems and related components, and is an essential foundation technology for next-generation core industries such as big data, smart factories, wearables, and smart homes. In particular, with the growth of cloud computing, the proportional expansion of the data center infrastructure is inevitable. Monitoring the health of these data center facilities is a way to maintain and manage the system and prevent failure. If a failure occurs in some elements of the facility, it may affect not only the relevant equipment but also other connected equipment, and may cause enormous damage. In particular, IT facilities are irregular due to interdependence and it is difficult to know the cause. In the previous study predicting failure in data center, failure was predicted by looking at a single server as a single state without assuming that the devices were mixed. Therefore, in this study, data center failures were classified into failures occurring inside the server (Outage A) and failures occurring outside the server (Outage B), and focused on analyzing complex failures occurring within the server. Server external failures include power, cooling, user errors, etc. Since such failures can be prevented in the early stages of data center facility construction, various solutions are being developed. On the other hand, the cause of the failure occurring in the server is difficult to determine, and adequate prevention has not yet been achieved. In particular, this is the reason why server failures do not occur singularly, cause other server failures, or receive something that causes failures from other servers. In other words, while the existing studies assumed that it was a single server that did not affect the servers and analyzed the failure, in this study, the failure occurred on the assumption that it had an effect between servers. In order to define the complex failure situation in the data center, failure history data for each equipment existing in the data center was used. There are four major failures considered in this study: Network Node Down, Server Down, Windows Activation Services Down, and Database Management System Service Down. The failures that occur for each device are sorted in chronological order, and when a failure occurs in a specific equipment, if a failure occurs in a specific equipment within 5 minutes from the time of occurrence, it is defined that the failure occurs simultaneously. After configuring the sequence for the devices that have failed at the same time, 5 devices that frequently occur simultaneously within the configured sequence were selected, and the case where the selected devices failed at the same time was confirmed through visualization. Since the server resource information collected for failure analysis is in units of time series and has flow, we used Long Short-term Memory (LSTM), a deep learning algorithm that can predict the next state through the previous state. In addition, unlike a single server, the Hierarchical Attention Network deep learning model structure was used in consideration of the fact that the level of multiple failures for each server is different. This algorithm is a method of increasing the prediction accuracy by giving weight to the server as the impact on the failure increases. The study began with defining the type of failure and selecting the analysis target. In the first experiment, the same collected data was assumed as a single server state and a multiple server state, and compared and analyzed. The second experiment improved the prediction accuracy in the case of a complex server by optimizing each server threshold. In the first experiment, which assumed each of a single server and multiple servers, in the case of a single server, it was predicted that three of the five servers did not have a failure even though the actual failure occurred. However, assuming multiple servers, all five servers were predicted to have failed. As a result of the experiment, the hypothesis that there is an effect between servers is proven. As a result of this study, it was confirmed that the prediction performance was superior when the multiple servers were assumed than when the single server was assumed. In particular, applying the Hierarchical Attention Network algorithm, assuming that the effects of each server will be different, played a role in improving the analysis effect. In addition, by applying a different threshold for each server, the prediction accuracy could be improved. This study showed that failures that are difficult to determine the cause can be predicted through historical data, and a model that can predict failures occurring in servers in data centers is presented. It is expected that the occurrence of disability can be prevented in advance using the results of this study.

• Korean Journal of Remote Sensing
• /
• v.28 no.4
• /
• pp.409-420
• /
• 2012

Since the Industrial Revolution, CO2 levels have been increasing with climate change. In this study, Analyze time-series changes in snow cover quantitatively and predict the vanishing point of snow cover statistically using remote sensing. The study area is Mt. Kilimanjaro, Tanzania. 23 image data of Landsat-5 TM and Landsat-7 ETM+, spanning the 27 years from June 1984 to July 2011, were acquired. For this study, first, atmospheric correction was performed on each image using the COST atmospheric correction model. Second, the snow cover area was extracted using the NDSI (Normalized Difference Snow Index) algorithm. Third, the minimum height of snow cover was determined using SRTM DEM. Finally, the vanishing point of snow cover was predicted using the trend line of a linear function. Analysis was divided using a total of 23 images and 17 images during the dry season. Results show that snow cover area decreased by approximately $6.47km^2$ from $9.01km^2$ to $2.54km^2$, equivalent to a 73% reduction. The minimum height of snow cover increased by approximately 290 m, from 4,603 m to 4,893 m. Using the trend line result shows that the snow cover area decreased by approximately $0.342km^2$ in the dry season and $0.421km^2$ overall each year. In contrast, the annual increase in the minimum height of snow cover was approximately 9.848 m in the dry season and 11.251 m overall. Based on this analysis of vanishing point, there will be no snow cover 2020 at 95% confidence interval. This study can be used to monitor global climate change by providing the change in snow cover area and reference data when studying this area or similar areas in future research.

• Journal of the Korean earth science society
• /
• v.22 no.5
• /
• pp.388-404
• /
• 2001

In order to examine the relative accuracy of satellite observations and model reanalyses about lower stratospheric temperature trends, two satellite-observed Microwave Sounding Unit (MSU) channel 4 (Ch 4) brightness temperature data and two GCM (ECMWF and GEOS) reanalyses during 1981${\sim}$1993 have been intercompared with the regression analysis of time series. The satellite data for the period of 1980${\sim}$1999 are MSU4 at nadir direction and SC4 at multiple scans, respectively, derived in this study and Spencer and Christy (1993). The MSU4 temperature over the globe during the above period shows the cooling trend of -0.35 K/decade, and the cooling over the global ocean is 1.2 times as much as that over the land. Lower stratospheric temperatures during the common period (1981${\sim}$1993) globally show the cooling in MSU4 (-0.14 K/decade), SC4 (-0.42 K/decade) and GEOS (-0.15 K/decade) which have strong annual cycles. However, ECMWF shows a little warming and weak annual cycle. The 95% confidence intervals of the lower stratospheric temperature trends are greater than those of midtropospheric (channel 2) trends, indicating less confidence in Ch 4. The lapse rate in the trend between the above two atmospheric layers is largest over the northern hemispheric land. MSU4 has low correlation with ECMWF over the globe, and high value with GEOS near the Korean peninsula. Lower correlations (r < 0.6) between MSU4 and SC4 (or ECMWF) occur over $30^{\circ}$N latitude belt, where subtropical jet stream passes. Temporal correlation among them over the globe is generally high (r > 0.6). Four kinds of lower stratospheric temperature data near the Korean peninsula commonly show cooling trends, of which the SC4 values (-0.82 K/decade) is the largest.

• Economic and Environmental Geology
• /
• v.34 no.4
• /
• pp.375-383
• /
• 2001

We extracted sea surface heights(SSH) from the TopexJPoseidon(T/P) radar altimeter data to compare with fhe SSH estimated from in-situ lide gauges(T/G) at Ulleungdo, Pohang, and SockcholMucko sites. Selection criteria such as wet/dry troposphere, ionosphere, and ocean tide were used to estimate accurate SSH. For time series analysis, the one-hour interval tide gauge SSHs were resampled al lO-day interval of the satellite SSHs. The ocean tide model applied in the altimeter data processing showed periodic aliasings of 175.5 day, 87.8 day, 62J day, 58.5 day, 49.5 day and 46.0 day, and, hence, the ZOO-day filtering was applied to reduce these spectral noises. Wavenumber correlation analysis was also applied to extract common components between the two SSHs, resulting in enhancing the correlation coefficient(CC) dramatically. The original CCs between the satenite and tide gauge SSHs are 0.46. 0.26, and 0.]5, respectively. Ulleungdo shows the largest cc bec;luase the site is far from the coast resulting in the minimun error in the satellite observations. The CCs were then increased to 0.59, 030, and 0.30, respectively, after 200.day filtering, and to 0.69, 0.63. and 0.59 after removing inversely correlative components using wavenumber correlation analysis. The CCs were greatly increased by 87, 227, and 460% when the wavenumber correlation analysis was followed by 2oo-day filtering, resulting in the final CCs of 0.86, 0.85, 0.84, respectively. It was found that the best SSHs were estimated when the two methods were applied to the original data. The low-pass filtered TIP SSHs were found to be well correlated with the TIG SSHs from tide gauges, and the best correlation results were found when we applied both low-pass filtering and spectral correlation analysis to the original SSHs.

• Journal of Soil and Groundwater Environment
• /
• v.13 no.4
• /
• pp.8-21
• /
• 2008

A series of three-dimensional numerical simulations using a hydrodynamic dispersion numerical model is performed to analyze quantitatively impacts of layered heterogeneity of geologic media and groundwater pumping schemes on groundwater flow and salt transport in coastal aquifer systems. A two-layer heterogeneous coastal aquifer system composed of a lower sand layer (aquifer) and an upper clay layer (aquitard) and a corresponding single-layer homogeneous coastal aquifer system composed of an equivalent lumped material are simulated to evaluate impacts of layered heterogeneity on seawater intrusion. In addition, a continuous groundwater pumping scheme and two different periodical groundwater pumping schemes, which withdraw the same amount of groundwater during the total simulation time, are applied to the above two coastal aquifer systems to evaluate impacts of groundwater pumping schemes on seawater intrusion. The results of the numerical simulations show that the periodical groundwater pumping schemes have more significant adverse influences on groundwater flow and salt transport not only in the lower sand layer but also in the upper clay layer, and groundwater salinization becomes more intensified spatially and temporally as the pumping intensity is higher under the periodical groundwater pumping schemes. These imply that the continuous groundwater pumping scheme may be more suitable to minimize groundwater salinization due to seawater intrusion. The results of the numerical simulations also show that groundwater salinization in the upper clay layer occurs significantly different from that in the lower sand layer under the periodical groundwater pumping schemes. Such differences in groundwater salinization between the two adjacent layers may result from layered heterogeneity of the layered coastal aquifer system.

• Journal of Korea Water Resources Association
• /
• v.53 no.2
• /
• pp.107-119
• /
• 2020

This study analyzed past drought characteristics based on the observed rainfall data and performed a long-term outlook for future extreme droughts using Representative Concentration Pathways 8.5 (RCP 8.5) climate change scenarios. Standardized Precipitation Index (SPI) used duration of 1, 3, 6, 9 and 12 months, a meteorological drought index, was applied for quantitative drought analysis. A single long-term time series was constructed by combining daily rainfall observation data and RCP scenario. The constructed data was used as SPI input factors for each different duration. For the analysis of meteorological drought observed relatively long-term since 1954 in Korea, 12 rainfall stations were selected and applied 10 general circulation models (GCM) at the same point. In order to analyze drought characteristics according to climate change, trend analysis and clustering were performed. For non-stationary frequency analysis using sampling technique, we adopted the technique DEMC that combines Bayesian-based differential evolution ("DE") and Markov chain Monte Carlo ("MCMC"). A non-stationary drought frequency analysis was used to derive Severity-Duration-Frequency (SDF) curves for the 12 locations. A quantitative outlook for future droughts was carried out by deriving SDF curves with long-term hydrologic data assuming non-stationarity, and by quantitatively identifying potential drought risks. As a result of performing cluster analysis to identify the spatial characteristics, it was analyzed that there is a high risk of drought in the future in Jeonju, Gwangju, Yeosun, Mokpo, and Chupyeongryeong except Jeju corresponding to Zone 1-2, 2, and 3-2. They could be efficiently utilized in future drought management policies.

• Journal of Korea Water Resources Association
• /
• v.49 no.2
• /
• pp.133-144
• /
• 2016

In this study, Frequency Domain Reflectometry (FDR) and COSMIC-ray soil moisture (SM) stations were installed at Sungkyunkwan University in Suwon, South Korea. To provide reliable information about SM, soil property test, time series analysis of measured soil moisture, and comparison of measured SM with satellite-based SM product are conducted. In 2014, six FDR stations were set up for obtaining SM. Each of the stations had four FDR sensors with soil depth from 5 cm to 40 cm at 5~10 cm different intervals. The result showed that study region had heterogeneous soil layer properties such as sand and loamy sand. The measured SM data showed strong coupling with precipitation. Furthermore, they had a high correlation coefficient and a low root mean square deviation (RMSD) as compared to the satellite-based SM products. After verifying the accuracy of the data in 2014, four FDR stations and one COSMIC-ray station were additionally installed to establish the Soil Moisture site with FDR and COSMIC-ray, called SM-FC. COSMIC-ray-based SM had a high correlation coefficient of 0.95 compared with mean SM of FDR stations. From these results, the SM-FC will give a valuable insight for researchers into investigate satellite- and model-based SM validation study in South Korea.

• Korean Journal of Environmental Agriculture
• /
• v.17 no.1
• /
• pp.70-75
• /
• 1998

A model of pond system is developed for treatment and recycling of excreta from twenty-five adult dairy cattle. It is composed of wastewater treatment ponds and small fish ponds. Those are three facultative ponds in series; primary-secondary-tertiary pond and these are designed to rear carps without feeding. A pit is constructed at the bottom of primary pond for efficient sludge sedimentation and effective methane fermentation. It is contrived to block into it the penetration of oxygen dissolved in the upper layer of pond water. The excreta from the cattle housed in stalls are diluted by water used for clearing them. The washed excreta flow into the pit. The average yearly $BOD_5$ concentration of influent is 398.7mg/l. That of the effluent from primary, secondary and tertiary pond of the system is 49.18, 27.9, and 19.8.mg/l respectively. Approximate 88, 93, and 95 % of BOD5 are removed in each pond. The mean yearly SS concentration of influent is 360.5 mg/l That of the effluent from each pond is 53.4, 45.7, and32.7mg/l respectively. Approximate 86, 88, and 91% of SS are removed in each pond. The $BOD_5$ concentration of secondary and tertiary pond can satisfy 30mg/l secondary treatment standard. The SS concentration of effluent from tertiary pond, however, is slightly greater than the standard, which results from activities of carps growing in the pond. The average yearly total nitrogen concentration of influent is 206.8mg/l and that of the effluent from each pond is 48.6, 30.8, and 21.0mg/l respectively. Approximate 74, 88, and 90% of total nitrogen are removed in each pond. The mean yearly total phosphorous concentration of influent is 20.7mg/l and that of the effluent from each pond is 5.3, 3.2, and 2.1mg/l respectively. Approximate 97, 98, and 99% of total phosphorous are removed in each pond. The high removal of nitrogen and phosphorous results from active growth of algae in the upper layer of pond water. Important pond design parameters for southern part of Korea -- areal loading of BOD5, liquid depth, hydraulic detention time, free board, and pond arrangement -- are taken up.

• KDI Journal of Economic Policy
• /
• v.31 no.1
• /
• pp.105-138
• /
• 2009

Although the decline in fertility rate is generally observed along the history of economic development throughout the world, the continuing decline hitting below the replacement level in Korea over the recent years gathered serious social concerns on the ground that it accelerates the process of population aging. The total fertility rate in Koreareached 2.08 in 1983, and gradually fell to the levels of 1.08 in 2005 and 1.26 in 2007. The policy debate over the role of the government has been focused mainly on the level of theoretical discussion without substantial basis on firm empirical evidence and the determinants of fertility. The objective of the paper is to empirically investigate the fertility effect of the female wage, which is understood as one of the most important determinants of fertility in Koreasince 1980 focusing on one aspect of fertility, namely birth spacing. Using the Korean National Fertility Survey conducted in 2006, I estimate a duration model of first and second births taking into account individual heterogeneity, which turned out to be an important factor to control for. Compared with previous studies in the literature on the Korean fertility, the study has an advantage of using the complete pregnancy history of women in a more representative sample. Unlike the previous studies, the analysis also deals with the endogeneity of marriage by treating a certain age, rather than age at marriage, as the time in which a woman becomes exposed to the risk of pregnancy. The study shares the common problem in the literature on birth spacing of lacking relevant wage information for respondents in a retrospective survey. I estimate the wage series as a function of the basic characteristics using the annual Wage Structure Survey from 1980 to 2005, which is considered as a nationally representative sample for wage information of employees. The results suggest that the increase in female wage by 10 percent leads to a decrease in second birth hazard by 0.56~0.92 percentage points and that the increase in spouse's wage by the equal amount is accompanied by the increase in second birth hazard by 0.36~1.13 percentage points. These estimates are more precisely estimated and of smaller magnitude than those presented by the previous studies. The results are robust to the different specifications of the wage equation. The simulation analysis based on the predicted values shows that about 17% of the change in the second birth hazard over the period 1980 to 2005 was due to the change in the female wage. Although there is some limitation in data, the results can be viewed as one estimate of the role of female wage on the recent fertility decline in Korea. The question raised by the paper is not a normative one of whether a government should promote childbearing but a positive one thatexplains fertility decline. Therefore, if there is a wide consensus on promoting childbearing, the finding suggests that the policies designed to reduce the opportunity cost of women in the labor market would be effective. The recent movement of implementing a wide range of family-friendly policies including child care support, maternity leave, parental leave and tax benefit in developed countries should be understood in this context.

• The Journal of Korean Society for Radiation Therapy
• /
• v.16 no.2
• /
• pp.69-79
• /
• 2004

In the process of photon treatments, linear accelerators with energies higher than 10 MV produce neutrons through the (${\gamma}$, n) interactions with the composite materials of the linac head md these materials further produce the induced radiations. We investigate the possible risks from these induced radiations especially in the wedge filters to the radiation workers. Wedge filters are used to modify the isodose profiles in the radiation treatment using the linear accelerator and always be handled by the radiation workers. For the background radiation, we measured the radiation in both the waiting room and the outside of the building for two hospitals, S and H. The results of S hospital were $0.11\;{\mu}Sv/hr$ and $0.10\;{\mu}Sv/hr$ for waiting room and outside respectively, and in the case of H hospital, they were $0.12\;{\mu}Sv/hr$ and $0.11\;{\mu}Sv/hr$. Using a survey meter, we measured the radiation from wedge filters inserted in 10 MV and 15 MV Siemens linear accelerators. The time series measurements were done in ${\sim}1$ minutes after exposure of 5 Gy of monitor units for the field size of $25{\times}25cm^2$. The starting value of 10 MV machine was about $3.26\;{\mu}Sv/hr$, which was three times higher than that of 10 MV. The measured radiation was from $^{28}Al$ and $^{53}Fe$ with a half life of 3.5 min. If the treatment patients are $20{\sim}50$ per day and the number of process of wedge filter change per patient is one or two, the annual dose equivalent is $0.08{\sim}0.4\;mSv$ for 10 MV, and $0.27{\sim}1.36\;mSv$ for 15 MV, which are in the range of dose equivalent limits of radiation workers.