• Korean Journal of Construction Engineering and Management
• /
• v.24 no.6
• /
• pp.24-35
• /
• 2023

Recently, the Ministry of Land, Infrastructure and Transport presented the '6th Construction Technology Promotion Basic Plan' and 'Smart Construction Revitalization Plan (2022.7.20)'. Off-Site Construction (OSC), which involves construction and production of PC (Precast Concrete) and Modular, etc., has advantages in shortening the construction period, reducing costs, improving quality, reducing construction waste, and reducing safety accidents. However, the construction cost is high compared to the traditional RC construction method, which has hindered its utilization and spread. In this study, OSC utilization was improved. An economic analysis indicator and methodology that can support decision-making in the planning and design stages for multi-unit housing were proposed. The factors used in the economic analysis of OSC (based on the PC method) of apartment houses were reviewed. As for the indicators used in the cost and benefit section, 'Construction Period', 'Disaster Occurrence', 'Waste Generation', and 'Greenhouse gas Emission', which reflect the technical advantages of OSC, were derived. In addition, a scenario analysis was conducted based on actual apartment housing case data for the presented economic analysis indicators and benefit calculation standards. The level of benefit that offsets the difference between the existing RC construction method and the construction cost was reviewed. In future studies, it will be necessary to conduct additional case studies to apply the measurement criteria for detailed indicators and supplement the benefit indicators.

• Clean Technology
• /
• v.30 no.1
• /
• pp.37-46
• /
• 2024

Environmental problems caused by plastic waste have been continuously growing around the world, and plastic waste is increasing even faster after COVID-19. In particular, PP and PE account for more than half of all plastic production, and the amount of waste from these two materials is at a serious level. As a result, researchers are searching for an alternative method to plastic recycling, and plastic pyrolysis is one such alternative. In this paper, a numerical study was conducted on the pyrolysis behavior of non-condensable gas to predict the chemical reaction behavior of the pyrolysis gas. Based on gas products estimated from preceding literature, the behavior of non-condensable gas was analyzed according to temperature and residence time. Numerical analysis showed that as the temperature and residence time increased, the production of H₂ and heavy hydrocarbons increased through the conversion of the non-condensable gas, and at the same time, the CH₄ and C₆H₆ species decreased by participating in the reaction. In addition, analysis of the production rate showed that the decomposition reaction of C₂H₄ was the dominant reaction for H₂ generation. Also, it was found that more H₂ was produced by PE with higher C₂H₄ contents. As a future work, an experiment is needed to confirm how to increase the conversion rate of H₂ and carbon in plastics through the various operating conditions derived from this study's numerical analysis results.

• Journal of Industrial Convergence
• /
• v.22 no.2
• /
• pp.73-83
• /
• 2024

The purpose of this study is to diagnose various problems arising around social enterprises in the sport field from the perspective of the organization and derive necessary tasks and implications. In order to achieve the purpose of the study, the study was largely divided into three stages, and the results were derived. First, the main status and characteristics of social enterprises in the sport field were examined. The current status was analyzed focusing on aspects such as background and origin, legislation and policy, organizational goals, organizational structure and procedures, and organizational characteristics. Social enterprises in the sport sector were in their early stages, and the government's social enterprise policy goal tended to focus on increasing the number of social enterprises in a short period of time through financial input. In addition, it was found that most individual companies rely on government subsidy support due to insufficient profit generation capacity. In the second stage, we focused on the situational factors that affect the functional performance of social enterprises in the sport field. As a result of reviewing the value, ideology, technology, and history of the organization, which are situational factors, it was derived that when certified as a social enterprise in the sport field and supported by the central government or local governments, political control is strong to some extent and exposure to the market is not severe. In the last third step, tasks and implications were derived to form an appropriate organization for social enterprises in the sport field. After the social enterprise ecosystem in the sport sector has been established to some extent, it is necessary to gradually move from the current "government-type" organization to the "national enterprise" organization. This is true in light of the government's limited financial level, not in the short term, but in order for the organization of social enterprises in the sports sector to survive in the long term.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.13
• /
• pp.1-40
• /
• 1973

These studies were aimed at clarification of genetic and ecological variation in culm length, panicle length and plant height of the $\textrm{F}_2$ plants in some selected crosses made between semi-dwarf rice varieties and tall Japonica ones. One Indica semi-dwarf, Taichung Native 1, one Indica $\times$ Japonica hybrid, IE51 and one Japonica semi-dwarf, Tankanbaekmang were used as short-gene donors while two of medium maturity varieties, Jinheung and Kwanok and one late veriety, Palkweng were used as the corresponding counterpart of respective dwarf varieties in a series of crosses. Five different crosses, Kwanok $\times$ Tankanbaekmang, Palkweng $\times$ Tankanbaekmang, Jinheung $\times$ T(N)1, Kwanok $\times$ T(N)1 and Kwanok $\times$ IE51, were made among the above six varieties. The $\textrm{F}_2$ plants of these crosses together with the concerned parental varieties were grown under several different conditions including three levels of each nitrogen and planting space, three planting seasons and three locations in 1968, to investigate variation in length of culm and panicle, and plant height. On the other hand, the F$_3$ progenies which were derived from the shortest 10 percent of the plants of three $\textrm{F}_2$ populations, Kwanok $\times$ T(N)1, Jinheung $\times$ T(N) 1 and Kwanok $\times$ IE51 grown in the previous year, were compared each other on the basis of selection efficiency in culm length. The experimental results could be summarized as follows; 1. Genetic behavior A. It was revealed that Tankanbaekmang, one of Japonica dwarf has a simple recessive gene responsible for short culm expression, showing a typical segregation ratio of three tall to one short culm plants in $\textrm{F}_2$ generation of the crosses either with Kwanok or Palkweng. B. In the both combinations, segregation pattern of the panicle length was exactly same as that of culm length. It seems that the same gene controls both culm length and panicle length. C. No difference between segregation of culm length and plant height in the above crosses was observed. D. T(N)1, one of Indica semi-dwarf did not show such a simple genetic behavior as detected from the crosses with Tankanbaekmang in segregation of culm length but formed a continuous and normal distribution curve. Therefore, some nonallelic genic actions might be involved in expression of culm length of the counterpart varieties of T(N)1. In particular, a transgressive segregation appeared toward the direction of longer culm length in case of Jinheung $\times$ T(N)1. The genetic behavior of panicle length and plant height generally coincided with that of culm length in all the cases. E. IE51 demonstrated exactly the same genetic behavior as that of T(N)1 when this variety was crossed with Kwanok. It was clearly clarified that the simple recessive gene controlling dwarfism from T(N)1 was well incorporated into this variety. 2. Ecological variation A. In general, there was a decreasing tendency in culm length and plant height of rice plant as seeding delayed while it was not so noticeable in panicle length. The decreasing magnitude varied from variety to variety and from cross to cross. Genetic behavior of the culm length and related characters of these materials was not disturbed by the variation of seeding season, nitrogen level, planting space and experimental location. E. The elongation mode of the upper three internodes was very similar to the segregation mode of culm length, panicle length and plant height in $\textrm{F}_2$ populations of . all the crosses investigated in this study. Accordingly, this result confirmed that the roles of the upper three internodes are very important in manifesting plant stature in rice. C. The effect of nitrogen on culm length and the related other two characters seemed to be meager. However, it was true to show an increasing tendency of those characters as nitrogen level got increased from 4 kg to 12kg per l0a, with different magnitude depending upon variety or cross. D. Also, the effect of planting space on culm length, panicle length and plant height was relatively small in all the cases. Those characters varied again depending upon variety or cross. However, a general increasing tendency was detected in manifestation of those traits under denser planting space condition. E. All the parental varieties produced shorter culm, panicle and plant height when they were grown at the lower latitude locations. It might be attributed to the fact that their reproductive growth accelerated with increased temperature prevailing at the lower latitude locations such as Iri and Mi1yang. On the countrary, $\textrm{F}_2$ population reacted differently to the different locations from the parental varieties. All the $\textrm{F}_2$ plants produced the longest culm, panicle and plant at Milyang. 3. Selection efficiency A. The heritability of culm length in Kwanok $\times$ T(N)1, Kwanok $\times$ IE51 and Jinheung$\times$T(N)1 was 92 percent, 74 percent and 55 percent, respectively. B. The actual genetic advance for culm length obtained from the progeny lines of the selected plants(10 precent) from the $\textrm{F}_2$ generation, was comparable to the expected advance calculated from the original $\textrm{F}_2$ populations. As compared with the $\textrm{F}_2$ population, the $\textrm{F}_3$ plants of Kwanok $\times$ T(N)l shortened on the average by 20.8cm, those of Kwanok $\times$ IE51 did 8.7cm and those of Jinheung$\times$T(N)1 20.0cm, respectively. C. Panicle length of the populations was differently affected from one cross to another by the selection based upon culm length in $\textrm{F}_2$ Kwanok $\times$ T(N)1 did not show any noticeable shortening of its culm length due to the selection pressure. On the other hand, both Kwanok $\times$ IE51 and Jinheung $\times$ T(N)1 showed a considerable shortening of their panicles in case of selection for culm length. Based upon the above results, it could be concluded that the ecological variation in culm length, panicle length and plant height was relatively small and fallen within the range of genetic variation. Considering from the fact that the simple recessive gene governing short height of Tankanbaekmang always accompanied with some undesirable characters such as short panicle and extremely small grain, the short gene of T(N)1 seemed to be more useful as dwarf gene source since it did not carry short gene together with such undesirable traits.

• 한국유가공학회:학술대회논문집
• /
• 2007.09a
• /
• pp.49-58
• /
• 2007

Inflammatory process leads to the well-known mucosal damage and therefore a further disturbance of the epithelial barrier function, resulting abnormal intestinal wall function, even further accelerating the inflammatory process[1]. Despite of the records, etiology and pathogenesis of IBD remain rather unclear. There are many studies over the past couple of years have led to great advanced in understanding the inflammatory bowel disease(IBD) and their underlying pathophysiologic mechanisms. From the current understanding, it is likely that chronic inflammation in IBD is due to aggressive cellular immune responses including increased serum concentrations of different cytokines. Therefore, targeted molecules can be specifically eliminated in their expression directly on the transcriptional level. Interesting therapeutic trials are expected against adhesion molecules and pro-inflammatory cytokines such as TNF-${\alpha}$. The future development of immune therapies in IBD therefore holds great promises for better treatment modalities of IBD but will also open important new insights into a further understanding of inflammation pathophysiology. Treatment of cytokine inhibitors such as Immunex(Enbrel) and J&J/Centocor(Remicade) which are mouse-derived monoclonal antibodies have been shown in several studies to modulate the symptoms of patients, however, theses TNF inhibitors also have an adverse effect immune-related problems and also are costly and must be administered by injection. Because of the eventual development of unwanted side effects, these two products are used in only a select patient population. The present study was performed to elucidate the ability of TNF-${\alpha}$ antibodies produced in sheep colostrums to neutralize TNF-${\alpha}$ action in a cell-based bioassay and in a small animal model of intestinal inflammation. In vitro study, inhibitory effect of anti-TNF-${\alpha}$ antibody from the sheep was determined by cell bioassay. The antibody from the sheep at 1 in 10,000 dilution was able to completely inhibit TNF-${\alpha}$ activity in the cell bioassay. The antibodies from the same sheep, but different milkings, exhibited some variability in inhibition of TNF-${\alpha}$ activity, but were all greater than the control sample. In vivo study, the degree of inflammation was severe to experiment, despite of the initial pilot trial, main trial 1 was unable to figure out of any effect of antibody to reduce the impact of PAF and LPS. Main rat trial 2 resulted no significant symptoms like characteristic acute diarrhea and weight loss of colitis. This study suggested that colostrums from sheep immunized against TNF-${\alpha}$ significantly inhibited TNF-${\alpha}$ bioactivity in the cell based assay. And the higher than anticipated variability in the two animal models precluded assessment of the ability of antibody to prevent TNF-${\alpha}$ induced intestinal damage in the intact animal. Further study will require to find out an alternative animal model, which is more acceptable to test anti-TNF-${\alpha}$ IgA therapy for reducing the impact of inflammation on gut dysfunction. And subsequent pre-clinical and clinical testing also need generation of more antibody as current supplies are low.

• Tuberculosis and Respiratory Diseases
• /
• v.50 no.4
• /
• pp.450-461
• /
• 2001

Background : The underlying pathogenesis of radiation-induced lung fibrosis (RTLF) has not been very well defined. However, the role of TGF-$\beta$ in the generation of RTLF has been a major focus because there is an increase in the expression of both the TGF-${\beta}m$-RNA and its protein preceding RTLF lesions. The down stream signal after a TGF-$\beta$ stimulated lung fibrosis includes the activation of many mediators such as Smad and c-Jun N-terminal kinase (JNK) through TAK1. It is we hypothesized that JNK activation may play a pivotal role in RTLF pathogenesis through increased transcription of the fibrogenic cytokines. The present study evaluates JNK activity in alveolar macrophages after irradiation and the relationship between JNK activity and the amount of collagen in the lung tissues. Methods : C57BL/6 mice(20-25 gr, males) received chlorotetracycline(2g/L) in their drinking water 1 week prior to irradiation and continuously there after. The mice were irradiated once with 1400 cGy of $60CO{\gamma}$-ray over the whole chest. The cellular composition of the whole lung bronchoalveoalr lavage fluids(BALF), elastin expression in the lung tissues, the level of hydroxyproline in lung tissues, and an in vitro JNK assay was measured before irradiation and one, four, and eight weeks after irradiation (RT). Results : The volumes of BALF retrieved from instilled 4 mL of saline with 2% heparin were 3.7-3.8 mL for each group. The cell numbers were similar before($4.1{\times}10^4{\pm}0.5{\times}10^4/mL$) and 1 week($3.1{\times}10^4{\pm}0.5{\times}10^4/mL$) after RT. At four and eight weeks after RT, the cell number reached to $14.0{\times}10^4{\pm}1.5{\times}10^4mL$ and $10.0{\times}10^4{\pm}1.3{\times}10^4/mL$, respectively. There we no changes in the lymphocytes and neutrophils population observed in the BALF after RT. The H-E stain of the lung tissues did not show any structural and fibrotic change in the lung tissues at 4 and 8 weeks after RT. In addition, the amount of elastin and collagen were not different on Verhoeff staining of the lung tissues before RT to eight weeks after RT. The hydroxyproine content was measured with the left lung dissected from the left main bronchus. The lung were homogenized and hydrolyzed with 6 N Hel for 12 hours at $110^{\circ}C$ then measured as previously described. The content of hydroxyproline, standardized with a lung protein concentration, reached a peak 4 weeks after RT, and thereafter showed a plateau. AnIn vitro JNK assay using c-$Jun_{1-79}$-GST sepharose beads were performed with the alveolar macrophages obtained from the BAL. JNK activity was not detected prior to RT, However, the JNK activity increased from one week after RT and reached a peak four weeks after RT. Conclusion : JNK may be involved in the pathogenesis because the JNK activity showed similar pattern observed with the hydroxyproine content. However, it is necessary to clarify that the JNK increases the transcription of fibrogenic cyiokines through the transcription factor.

• Journal of Global Scholars of Marketing Science
• /
• v.18 no.1
• /
• pp.59-80
• /
• 2008

Recently, e-sports are growing with potentiality as a new industry with conspicuous profit model. But studies that dealing with e-sports are not enough. Hence, proposes of this paper are both to establish basic model that is for the design of e-sport marketing strategy and to contribute toward future studies which are related to e-sports. Recently, the researches to explain sports-sponsorship through the identification theory have been discovered. Many researches say that somewhat proper identification is a requirement for most sponsors to improve the their images which is essential to sponsorship activity. Consequently, the research for sponsorship associated with identification in the e-sports, not in the physical sports is the core sector of this study. We extracted the variables from online's major characteristics and existing sport sponsorship researches. First, because e-sports mean the tournaments or leagues in the use of online game, the main event of the game is likely to call it online game. Online media's attributes are distinguished from those of offline. Especially, interactivity, anonymity, and expandibility as a e-sport game attributes are able to be mentioned. So, these inherent online attributes are examined on the relationship with flow. Second, in physical sports games, Fisher(1998) revealed that team similarity and team attractivity were positively related to team identification. Wann(1996) said that the result of former game influenced the evaluation of the next game, then in turn has an effect on the identification of team supporters. Considering these results in the e-sports side, e-sports gamer' attractivity, similarity, and match result seem to be important precedent variables of the identification with a gamer. So, these e-sport gamer attributes are examined on the relationship with both flow and identification with a gamer. Csikszentmihalyi(1988) defined the term flow as feeling status for him to be making current positive experience optimally. Hoffman and Novak(1996) also said that if a user experienced the flow he would visit a website without any reward. Therefore flow might be positively associated with user's identification with a gamer. And, Swanson(2003) disclosed that team identification influenced the positive results of sponsorship, which included attitude toward sponsors, sponsor patronage, and satisfaction with sponsors. That is, identification with a gamer expect to be connected with corporation identification significantly. According to the above, we can design the following research model. All variables used in this study(interactivity, anonymity, expandibility, attractivity, similarity, match result, flow, identification with a gamer, and identification with a sponsor) definitely were defined operationally underlying precedent researches. Sample collection was carried out to the person who has an experience to have enjoyed e-sports during June 2006. Much portion of samples is men because much more men than women enjoy e-sports in general. Two-step approach was used to test the hypotheses. First, confirmatory factor analysis was committed to guarantee the validity and reliability of variables. The results showed that all variables had not only intensive and discriminant validity, but also reliability. Then, research model was examined with fully structural equation using LISREL 8.3 version. The fitness of the suggested model mostly was at the acceptable level. Shortly speaking about the results, first of all, in e-sports game attributes, only interactivity which is called a basic feature in online situation affected flow positively. Secondly, in e-sports gamer's attributes, similarity with a gamer and match result influenced flow positively, but there was no significant effect in the relationship between the attractivity of a gamer and flow. And as expected, similarity had an effect on identification with a gamer significantly. But unexpectedly attractivity and match result did not influence identification with a gamer significantly. Just the same as the fact verified in the many precedent researches, flow greatly influenced identification with a gamer, and identification with a gamer continually had an influence on the identification with a sponsor significantly. There are some implications in these results. If the sponsor of e-sports supports the pro-game player who absolutely should have the superior ability to others and is similar to the user enjoying e-sports, many amateur gamers will feel much of the flow and identification with a pro-gamer, and then after all, feel the identification with a sponsor. Such identification with a sponsor leads people enjoying e-sports to have purchasing intention for products produced by the sponsor and to make a positive word-of-mouth for those products or the sponsor. For the future studies, we recommend a few ideas. Based on the results of this study, it is necessary to find new variables relating to the e-sports, which is not mentioned in this study. For this work to be possible, qualitative research seems to be needed to consider the inherent e-sport attributes. Finally, to generalize the results related to e-sports, a wide range of generations not a specific generation should be researched.

• Journal of Intelligence and Information Systems
• /
• v.20 no.1
• /
• pp.163-176
• /
• 2014

Social media is becoming the platform for users to communicate their activities, status, emotions, and experiences to other people. In recent years, microblogs, such as Twitter, have gained in popularity because of its ease of use, speed, and reach. Compared to a conventional web blog, a microblog lowers users' efforts and investment for content generation by recommending shorter posts. There has been a lot research into capturing the social phenomena and analyzing the chatter of microblogs. However, measuring television ratings has been given little attention so far. Currently, the most common method to measure TV ratings uses an electronic metering device installed in a small number of sampled households. Microblogs allow users to post short messages, share daily updates, and conveniently keep in touch. In a similar way, microblog users are interacting with each other while watching television or movies, or visiting a new place. In order to measure TV ratings, some features are significant during certain hours of the day, or days of the week, whereas these same features are meaningless during other time periods. Thus, the importance of features can change during the day, and a model capturing the time sensitive relevance is required to estimate TV ratings. Therefore, modeling time-related characteristics of features should be a key when measuring the TV ratings through microblogs. We show that capturing time-dependency of features in measuring TV ratings is vitally necessary for improving their accuracy. To explore the relationship between the content of microblogs and TV ratings, we collected Twitter data using the Get Search component of the Twitter REST API from January 2013 to October 2013. There are about 300 thousand posts in our data set for the experiment. After excluding data such as adverting or promoted tweets, we selected 149 thousand tweets for analysis. The number of tweets reaches its maximum level on the broadcasting day and increases rapidly around the broadcasting time. This result is stems from the characteristics of the public channel, which broadcasts the program at the predetermined time. From our analysis, we find that count-based features such as the number of tweets or retweets have a low correlation with TV ratings. This result implies that a simple tweet rate does not reflect the satisfaction or response to the TV programs. Content-based features extracted from the content of tweets have a relatively high correlation with TV ratings. Further, some emoticons or newly coined words that are not tagged in the morpheme extraction process have a strong relationship with TV ratings. We find that there is a time-dependency in the correlation of features between the before and after broadcasting time. Since the TV program is broadcast at the predetermined time regularly, users post tweets expressing their expectation for the program or disappointment over not being able to watch the program. The highly correlated features before the broadcast are different from the features after broadcasting. This result explains that the relevance of words with TV programs can change according to the time of the tweets. Among the 336 words that fulfill the minimum requirements for candidate features, 145 words have the highest correlation before the broadcasting time, whereas 68 words reach the highest correlation after broadcasting. Interestingly, some words that express the impossibility of watching the program show a high relevance, despite containing a negative meaning. Understanding the time-dependency of features can be helpful in improving the accuracy of TV ratings measurement. This research contributes a basis to estimate the response to or satisfaction with the broadcasted programs using the time dependency of words in Twitter chatter. More research is needed to refine the methodology for predicting or measuring TV ratings.

• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.79-104
• /
• 2020

Recently, as deep learning has attracted attention, the use of deep learning is being considered as a method for solving problems in various fields. In particular, deep learning is known to have excellent performance when applied to applying unstructured data such as text, sound and images, and many studies have proven its effectiveness. Owing to the remarkable development of text and image deep learning technology, interests in image captioning technology and its application is rapidly increasing. Image captioning is a technique that automatically generates relevant captions for a given image by handling both image comprehension and text generation simultaneously. In spite of the high entry barrier of image captioning that analysts should be able to process both image and text data, image captioning has established itself as one of the key fields in the A.I. research owing to its various applicability. In addition, many researches have been conducted to improve the performance of image captioning in various aspects. Recent researches attempt to create advanced captions that can not only describe an image accurately, but also convey the information contained in the image more sophisticatedly. Despite many recent efforts to improve the performance of image captioning, it is difficult to find any researches to interpret images from the perspective of domain experts in each field not from the perspective of the general public. Even for the same image, the part of interests may differ according to the professional field of the person who has encountered the image. Moreover, the way of interpreting and expressing the image also differs according to the level of expertise. The public tends to recognize the image from a holistic and general perspective, that is, from the perspective of identifying the image's constituent objects and their relationships. On the contrary, the domain experts tend to recognize the image by focusing on some specific elements necessary to interpret the given image based on their expertise. It implies that meaningful parts of an image are mutually different depending on viewers' perspective even for the same image. So, image captioning needs to implement this phenomenon. Therefore, in this study, we propose a method to generate captions specialized in each domain for the image by utilizing the expertise of experts in the corresponding domain. Specifically, after performing pre-training on a large amount of general data, the expertise in the field is transplanted through transfer-learning with a small amount of expertise data. However, simple adaption of transfer learning using expertise data may invoke another type of problems. Simultaneous learning with captions of various characteristics may invoke so-called 'inter-observation interference' problem, which make it difficult to perform pure learning of each characteristic point of view. For learning with vast amount of data, most of this interference is self-purified and has little impact on learning results. On the contrary, in the case of fine-tuning where learning is performed on a small amount of data, the impact of such interference on learning can be relatively large. To solve this problem, therefore, we propose a novel 'Character-Independent Transfer-learning' that performs transfer learning independently for each character. In order to confirm the feasibility of the proposed methodology, we performed experiments utilizing the results of pre-training on MSCOCO dataset which is comprised of 120,000 images and about 600,000 general captions. Additionally, according to the advice of an art therapist, about 300 pairs of 'image / expertise captions' were created, and the data was used for the experiments of expertise transplantation. As a result of the experiment, it was confirmed that the caption generated according to the proposed methodology generates captions from the perspective of implanted expertise whereas the caption generated through learning on general data contains a number of contents irrelevant to expertise interpretation. In this paper, we propose a novel approach of specialized image interpretation. To achieve this goal, we present a method to use transfer learning and generate captions specialized in the specific domain. In the future, by applying the proposed methodology to expertise transplant in various fields, we expected that many researches will be actively conducted to solve the problem of lack of expertise data and to improve performance of image captioning.

• 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.