, pp.455-459 http://dx.doi.org/10.14257/astl.2016.139.90 Big Data Service Combination for Efficient Energy Data Analytics Tai-Yeon Ku, Wan-ki Park, Il-Woo Lee Energy IT Technology Research Section Hyper-connected Communication Research Laboratory, ETRI Dae-jeon, Korea kutai@etri.re.kr, wkpark@etri.re.kr, ilwoo@etri.re.kr Abstract. The paper relates to a system and method for combining a Big Data service for efficient energy data analytics. Big data platform technology has been developed, and various methods for utilizing the technology are being proposed, but only information technologies (ITs) for advanced users have been offered. This paper is directed to a system and method for combining a service combination capable of providing a big data analysis more easily and rapidly and recognizing an analysis flow more easily by combining a model-based service combination. The Big Data Service combination tool is to provide a system and a method for data analysis services which is able to select data corresponding to a user requirement and an analysis algorithm which is able to analyze the data and let an analysis service selected by the user among a plurality of analysis services be automatically performed in a big data platform. Keywords: Big Data, Analytics, Service Combination 1 Introduction Recently, a data market, an analysis algorithm market is based on the Hadoop. Many big data platform techniques have been tried to interconnect between data and an analysis algorithm but it is not still visualized since there is no standard method yet. Many data analysis algorithms and methods have been introduced technically but data owners have to be familiar with such intricate analysis methods or do not know how to utilize it. As a result, such valuable data becomes disappeared. Thus, various methods for utilizing data have been introduced and big data platform technics have been developed to resolve such problems but they are still only for professional users. 2 Overview of Big-Data Service Combination architecture Accordingly, the workflow model may be configured in a selection tree form including various workflow models according to a user s intention. That is, a plurality of workflow models in which each node configures an upper and lower level may be included in one workflow model. ISSN: 2287-1233 ASTL Copyright 2016 SERSC
Fig.1. Platform Independent Service Combination 3 Platform Independent Bigdata Service Combination Mechanism Fig. 2 is a diagram for describing a combination of a hierarchical model-based service combination. The levels of the service combination model may be generated differently according to which classification standard the service combination model is classified based on. For example, a classification according to a data type may be arranged in a first level, a classification according to an analysis domain may be arranged in a second level which is a lower level than the first level, and a classification according to a function may be arranged in a third level which is a lower level than the second level. 456 Copyright 2016 SERSC
Fig.2. Multi-Level Service Combination Model Accordingly, the service combination model may be configured in a selection tree form including various service combination models according to a user s intention. That is, a plurality of service combination models in which each node configures an upper and lower level may be included in one service combination model. When a selection input of a user for the classification standard is received and the selection tree form is configured as shown in FIG. 2, a service combination model P1 formed as one or more levels may be mapped to one or more nodes P2 configuring the service combination model. That is, one or more service combination models including their lower levels may be mapped to one node. The service combination generation tool may provide the service combination model to the user, and the user may select and utilize a service combination model P1 which the user wants. FIG. 3 is a diagram for describing a hierarchical structure of a service combination model. Fig.3. A hierarchical structure of a service combination Copyright 2016 SERSC 457
One service combination model P1 may be generated through various manners. For example, the service combination model P1 may be generated by receiving information corresponding to a node of a lower level input by a user, and may be generated through a combination of a service combination model which is previously generated. In order to generate the service combination model in the various manners, the service combination model P1 may have a multi-level structure. That is, a service combination model P1 in which a lower level for an advanced developer, a level providing a basic parameter value, a level of abstractly componentizing frequently used functions and providing the componentized functions, a level of automatically providing complex big data through designation of a simple input and output value, etc. are configured in the multi-level may be provided. In the method for combining a service combination, first, a classification standard for each level of the service combination model may be selected in response to a selection input of the user for the classification standard of the service combination model. Next, according to the classification standard for each level, one or more service combination models may be mapped to a node included in the service combination model. In this case, one classification standard may be applied to each level, and a service combination model formed as one or more levels may be mapped to the node. Meanwhile, each level of the service combination model may have a level different from other levels, and each level may include different authority and function. Next, the mapped service combination model may be displayed. Meanwhile, the method for combining the service combination may further include an operation of receiving a selection input for a node of each level of the mapped service combination model input by the user. That is, as the user selects any one node for each level, the node for each level which is selected may be displayed on the service combination generation space of the service combination GUI tool based on the sequence of the levels. In the description described above, according to implementation examples of the present paper, the operations may be separated to have further operations or be combined to have the smaller number of operations. Further, some operations may be omitted according to necessity, and the sequence of the operations may be changed. Moreover, even though some description have been omitted, the content described above regarding the proposed service combination mechanism for combining the service combination may also be applied to the method for combining the service combination. A plurality of users may easily and rapidly generate an analysis service through a combination of a puzzle form based on a defined model by defining ecosystems as the model and providing a reusable structure. Further, information such as a state of work which is currently being performed, work which is completed, and work which is not processed, etc. among work listed on a service combination may be easily recognized. Moreover, an analysis result may be provided in various manners through a visualization model. In addition, a setting error, etc. needed when generating or performing a service of a complex ecosystem through the provided service combination generation tool may be automatically detected, and it may be possible to generate a service using the minimum necessary input through a default setting value. Also, when updating the ecosystem, the previously generated service combination may be easily reusable through a model update of a related ecosystem, and thus it is possible to rapidly adapt to big data technology which is rapidly changing. Through this paper, big data analysis which has 458 Copyright 2016 SERSC
been performed by some specialists may be easily and rapidly provided, and an analysis flow may be easily recognized by a combination of the model-based service combination. Moreover, an analysis result may be provided in various manners through the visualization model. 4 Conclusion The paper relates to a system and method for combining a Big Data service for efficient energy data analytics. Big data platform technology has been developed, and various methods for utilizing the technology are being proposed, but only information technologies (ITs) for advanced users have been offered. This paper is directed to a system and method for combining a service combination capable of providing a big data analysis more easily and rapidly and recognizing an analysis flow more easily by combining a model-based service combination. Acknowledgements. This work was supported by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (No. 20161210200360). References 1. Ku, T.-y., Won, H.-S., Choi, H.: Service Recommendation System for Big Data Analysis, ICOIN 2016, Oct, 2015 2. Ku, T.-y., Won, H.-S., Choi, H.:, Adaptive Cache Deploying Architecture Using Big Data Framework for CDN, ICTC 2015, Oct, 2015 Copyright 2016 SERSC 459