• 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.

• Journal of Chest Surgery
• /
• v.41 no.3
• /
• pp.329-334
• /
• 2008

Background: Percutaneous cardiopulmonary support. (PCPS) has the potential to rescue patients in cardiogenic shock who might otherwise die. PCPS has been a therapeutic option in a variety of the clinical settings such as for patients with myocardial Infarction, high-risk coronary intervention and postcardiotomy cardiogenic shock, and the PCPS device is easy to install. We report our early experience with PCPS as a life saving procedure in cardiogenic shock patients due to acute myocardial infarction. Material and Method: From January 2005 to December 2006, eight patients in cardiogenic shock with acute myocardial infarction underwent PCPS using the CAPIOX emergency bypass system($EBS^{(R)}$, Terumo, Tokyo, Japan). Uptake cannulae were inserted deep into the femoral vein up to the right atrium and return cannulae were inserted into the femoral artery with Seldinger techniques using 20 and 16-French cannulae, respectively. Simultaneously, autopriming was performed at the $EBS^{(R)}$ circuit. The $EBS^{(R)}$ flow rate was maintained between $2.5{\sim}3.0L/min/m^2$ and anticoagulation was performed using intravenous heparin with an ACT level above 200 seconds. Result: The mean age of patients was $61.1{\pm}14.2$ years (range, 39 to 77 years). Three patients were under control of the $EBS^{(R)}$ before percutaneous coronary intervention (PCI), three patients were under control of the $EBS^{(R)}$ during PCI, one patient was under control of the $EBS^{(R)}$ after PCI, and one patient was under control of the $EBS^{(R)}$ after coronary bypass surgery. The mean support time was $47.5{\pm}27.9$ hours (range, 8 to 76 hours). Five patients (62.5%) could be weaned from the $EBS^{(R)}$ after $53.6{\pm}27.2$ hours. (range, 12 to 68 hours) of support. All of the patients who could successfully be weaned from support were discharged from the hospital. There were three complications: one case of gastrointestinal bleeding and two cases of acute renal failure. Two of the three mortality cases were under cardiac arrest before $EBS^{(R)}$ support, and one patient had an intractable ventricular arrhythmia during the support. All of the discharged patients are still surviving at $16.8{\pm}3.1$ months (range, 12 to 20 months) of follow-up. Conclusion: The use of $EBS^{(R)}$ for cardiogenic shock caused by an acute myocardial infarction could rescue patients who might otherwise have died. Successfully recovered patients after $EBS^{(R)}$ treatment have survived without severe complications. More experience and additional clinical investigations are necessary to elucidate the proper installation timing and management protocol of the $EBS^{(R)}$ in the future.

• Journal of Chest Surgery
• /
• v.38 no.8 s.253
• /
• pp.538-544
• /
• 2005

Background: Pulmonary artery banding (PAB) is an initial palliative procedure for a diverse group of patients with congenital cardiac anomalies and unrestricted pulmonary blood flow. We proved the usefulness of PAB through retrospective investigation of the surgical indication and risk analysis retrospectively. Material and Method: One hundred and fifty four consecutive patients (99 males and 55 females) who underwent PAB between January 1986 and December 2003 were included. We analysed the risk factors for early mortality and actuarial survival rate. Mean age was $2.5\pm12.8\;(0.2\sim92.7)$ months and mean weight was $4.5\pm2.7\;(0.9\sim18.0)\;kg$. Preoperative diagnosis included functional single ventricle $(88,\;57.1\%)$, double outlet right ventricle $(22,\;14.2\%)$, transposition of the great arteries $(26,\;16.8\%)$, and atrioventricular septal defect $(11,\;7.1\%)$. Coarctation of the aorta or interrupted aortic arch $(32,\;20.7\%)$, subaortic stenosis $(13,\;8.4\%)$ and total anomalous pulmonary venous connection $(13,\;8.4\%)$ were associated. Result: The overall early mortality was $22.1\%\;(34\;of\;154)$, The recent series from 1996 include patients with lower age $(3.8\pm15.9\;vs.\;1.5\pm12.7,\;p=0.04)$ and lower body weight $(4.8\pm3.1\;vs.\;4.0\pm2.7,\;p=0.02)$. The early mortality was lower in the recent group $(17.5\%;\;16/75)$ than the earlier group $(28.5\%;\;18/45)$. Aortic arch anomaly (p=0.004), subaortic stenosis (p=0.004), operation for subaortic stenosis (p=0.007), and cardiopulmonary bypass (p=0.007) were proven to be risk factors for early death in univariate analysis, while time of surgery (<1996) (p=0.026) was the only significant risk factor in multivariate analysis. The mean time interval from PAB to the second-stage operation was $12.8\pm10.9$ months. Among 96 patients who survived PAB, 40 patients completed Fontan operation, 21 patients underwent bidirectional cavopulmonary shunt, and 35 patients underwent biventricular repair including 25 arterial switch operations. Median follow-up was $40.1\pm48.9$ months. Overall survival rates at 1 year, 5 years and 10 years were $81.2\%\;65.0\%,\;and\;63.5\%$ respectively. Conclusion: Although it improved in recent series, early mortality was still high despite the advances in perioperative management. As for conventional indications, early primary repair may be more beneficial. However, PA banding still has a role in the initial palliative step in selective groups.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.21 no.4
• /
• pp.50-63
• /
• 2018

The rapid urbanization had led to a distortion of natural hydrological cycle system. The change in hydrological cycle structure is causing streamflow depletion, changing the existing use tendency of water resources. To manage such phenomena, a streamflow depletion impact assessment technology to forecast depletion is required. For performing such technology, it is indispensable to build GIS-based spatial data as fundamental data, but there is a shortage of related research. Therefore, this study was conducted to use the use of GIS-based time series spatial data for streamflow depletion assessment. For this study, GIS data over decades of changes on a national scale were constructed, targeting 6 streamflow depletion impact factors (weather, soil depth, forest density, road network, groundwater usage and landuse) and the data were used as the basic data for the operation of continuous hydrologic model. Focusing on these impact factors, the causes for streamflow depletion were analyzed depending on time series. Then, using distributed continuous hydrologic model based DrySAT, annual runoff of each streamflow depletion impact factor was measured and depletion assessment was conducted. As a result, the default value of annual runoff was measured at 977.9mm under the given weather condition without considering other factors. When considering the decrease in soil depth, the increase in forest density, road development, and groundwater usage, along with the change in land use and development, and annual runoff were measured at 1,003.5mm, 942.1mm, 961.9mm, 915.5mm, and 1003.7mm, respectively. The results showed that the major causes of the streaflow depletion were lowered soil depth to decrease the infiltration volume and surface runoff thereby decreasing streamflow; the increased forest density to decrease surface runoff; the increased road network to decrease the sub-surface flow; the increased groundwater use from undiscriminated development to decrease the baseflow; increased impervious areas to increase surface runoff. Also, each standard watershed depending on the grade of depletion was indicated, based on the definition of streamflow depletion and the range of grade. Considering the weather, the decrease in soil depth, the increase in forest density, road development, and groundwater usage, and the change in land use and development, the grade of depletion were 2.1, 2.2, 2.5, 2.3, 2.8, 2.2, respectively. Among the five streamflow depletion impact factors except rainfall condition, the change in groundwater usage showed the biggest influence on depletion, followed by the change in forest density, road construction, land use, and soil depth. In conclusion, it is anticipated that a national streamflow depletion assessment system to be develop in the future would provide customized depletion management and prevention plans based on the system assessment results regarding future data changes of the six streamflow depletion impact factors and the prospect of depletion progress.