Earticle

클라우드 컴퓨팅 기술의 활성화에 따라 데이터 센터도 클라우드화 되고 있다. 클라우드 데이터 센터는 유연한 비즈니스 수요에 대처하기 위해서 가상화 기술을 활용함으로써 기존의 데이터 센터 보다 ICT 자원을 효율적으로 이용할수 있다. 클라우드 데이터 센터는 ICT 장비의 고밀도 구성으로 인해 에너지 소비가 더 높아지게 되어 보다 유연한관리가 요구된다. 본 연구는 이러한 특징을 갖는 클라우드 데이터 센터의 에너지 효율성을 진단하기 위해서 인프라분야, ICT 장비 분야, 그리고 운영 및 에너지 효율 관리 분야 측면에서의 평가 프레임워크를 제시한다.

As cloud computing technology becomes popular, many data centers would like to evolve thetechnology foundation from legacy to cloud technology. A cloud technology based data center canmanage ICT resources more efficiently than traditional data center by utilizing virtualizationtechnology to deliver better flexibility on demand. However, a cloud data center needs to embracemore equipments with higher density, which consumes more energy, while it needs to provide moreflexible management. This research provides a framework to evaluate energy efficiency of clouddata center with categories of infrastructure, ICT equipment, and operation and energy efficiency.

본 연구는 지방소멸 시대 원도심 재생을 생활 기반 회복 관점에서 재구성하고, 공주시 원도심과 인천시 개항장 일대를 사례로 문헌과 공개자료 기반 비교사례연구를 수행하였다. 쇠퇴 메커니즘, 재생 지향점, 추진경로, 기대 성과 등의 동일 기준으로 두 사례를 비교하였다. 또한 이러한 비교를 바탕으로 원도심 재생을 정주 및 생활 기반형 재생과 관광 브랜딩형 재생으로 분류하였다. 최종 분석 결과 두 사례는 중심성 이동과 기능 재배치라는 구조적 조건을 공유하나 공주는 공공 기반 하에 활동 주체와 소규모 공간 운영이 축적되며 일상 기능과 정주 기반으로 확장될 여지가 상대적으로 큰 반면, 인천은 역사문화 자원을 매개로 방문 수요와 결합되는 관광 브랜딩형 경로가 전면화되는 차이가 확인되었다. 이를 바탕으로 본 연구는 경로별 성과와 리스크를 점검할 평가 프레임을 제안하고, 생활 기반 연결성, 상권 안정성, 이용 충돌 관리, 추진체계 지속성, 정체성 축적을 핵심 평가 영역으로 제시한다. 그리고 이러한 지표들을 통해 추후 도시재생의 성과를 계량적으로 측정할 수 있는 토대를 마련했다는 의의가 있다.

This study reframes downtown regeneration in an era of population decline from the perspective of restoring everyday living foundations. Using Gongju’s old downtown and Incheon’s Open Port area as cases, it conducts a comparative case study based on prior research and publicly available materials. The two cases are compared using a common set of criteria, including decline mechanisms, regeneration objectives, implementation pathways, and expected outcomes. On this basis, the study classifies downtown regeneration into two pathways: a residential and everyday-life-based pathway and a tourism-and-branding-oriented pathway. The analysis finds that both cases share structural conditions characterized by shifts in urban centrality and functional restructuring. However, Gongju shows greater potential for expansion toward everyday functions and residential foundations through the accumulation of local actors and small-scale space operations on a public-sector base, whereas Incheon demonstrates a more fully developed tourism-and-branding pathway that is strongly linked to visitor demand mediated by historical and cultural assets. Building on these findings, the study proposes an evaluation framework to assess pathway-specific outcomes and risks, and identifies five core evaluation domains: linkage to everyday living foundations, commercial district stability, management of use conflicts, sustainability of the implementation structure, and the accumulation of local identity. By specifying these indicators, the study also provides a foundation for future quantitative measurement of regeneration performance.

학습자가 맥락적 상황 속에서 개인 사회적 문제를 해결하기 위하여 자기주도적 탐구 및 창조 활동을 하며 가시적 결과물을 제작하는 것이 목적인 메이커 교육은 메이커 운동이 지닌 교육적 가치와 의미를 강조하고 있다. 즉, '메이커 정신'의 함양을 강조하며 기존의 교육과 다른 가치와 목적를 추구하므로 메이커 교육의 평가는 기존의 결과 중심의 평가가 아닌 과정 중심의 평가로 메이커 활동의 특성을 반영할 수 있어야 한다. 이러한 맥락에서 본 연구는 메이커 교육을 위한 새로운 평가틀 탐색을 위하여 문헌연구를 실시하고, 기존의 지 덕 체의 3 요소 외에 상호작용성 및 사회적 참여를 평가틀의 5가지 구성요소로 제안하였다. 곧, 5 가지 구성요소는 5 ONs(Minds-on, Hands-on, Hearts-on, Social-on, Acts-on)로 요약되었으며, 이와 더불어 각 요소별 세부항목을 규명하였다. 이후 이렇게 도출된 5 ONs의 구성요소 및 각 요소별 세부항목에 대한 내용 타당도를 위한 델파이 조사를 실시하였다. 그리고 그 결과로서 '5 ONs 기반의 평가틀 구성요소 및 세부항목'으로 구성된 메이커교육을 위한 평가틀을 제시하였다.

Maker education rooted on Maker Movement refers to constructivist learning approach in which students as makers participate in producing visible outcomes through self-directed inquiry and creative hands-on activities in a real life context to solve their personal or social problems. The Maker education, therefore, stresses cultivation of 'maker mindset' in the process-oriented learning environments, pursuing evaluation aspects different from those in the existing educational system. In this context, this study aimed to explore an evaluation framework for the Maker education which reflects the Maker mindset: First, a literature review was conducted to search for the evaluation framework of the maker education which consists of the category of 5 ONs (Minds-on, Hands-on, Hearts-on, Social-on, Acts-on) representing intellectual, physical, emotional, interpersonal and practical aspects, respectively; Second, a Delphi survey for content validity was carried out to confirm the adequacy of the 5 ONs category along with sub-elements for each category. Finally, this study presented the evaluation framework for the Maker education, which is expected to be used as feedback rather than a measuring tool for the process and environments of the Maker education.

Topic modeling has evolved from statistical methods such as latent Dirichlet allocation (LDA) to neural hybrid models including BERTopic, which utilize bidirectional encoder representations from transformers (BERT) embeddings. However, traditional statistical evaluation metrics overlook the semantic richness of these neural representations, limiting model assessment capabilities. This paper introduces semantic-based evaluation metrics that leverage deep learning embeddings and validates them through both statistical comparison and large language model (LLM)-based assessment. This study evaluated three synthetic datasets with systematically varying topic overlap and one public dataset (20 Newsgroups). Analysis across 9,608 synthetic documents with 45 topics and a stratified sample of 1,000 documents from 20 Newsgroups shows that semantic metrics achieve improved discrimination compared to statistical baselines. Specifically, semantic coherence shows a 38.1% discriminative range versus 5.0% for statistical measures, representing a 7.62× improvement. Semantic distinctiveness achieves 1.57× higher discrimination than statistical methods. Semantic methods also maintain consistent discrimination quality for diversity metrics, with stable progression across similarity levels. LLM assessments, serving as proxies for human judgment, demonstrate inter-model agreement through a weighted three-model ensemble (mean pairwise Spearman ρ=0.937) and positive correlation with semantic metrics on public datasets (ρ=0.632-0.671). Domain-specific validation and multilingual extension constitute future work.

Recently, the development of sustainable building assessment tools as means to invigorate the dissemination of sustainable buildings has been actively progressed. However, many assessment tools involve various problems in terms of assessment method and system framework, which greatly impede their credibility and applicability. If these problems persist over time, the role of sustainable building assessment tools as decision making measures during the design stage will be greatly limited. The objective of the study is to suggest a systematic model for sus- tainable building assessment tools by establishing a logical system of performance assessment framework. For this purpose, the Environmen- tal Impact Assessment(EIA) framework used in selected and modified to fit the building performance assessment. The analysis of performance assessment tools for sustainable buildings was conducted using the EIA framework. Based on the results of the analysis, a framework for the performance assessment of sustainable buildings was established.

Background: The internal processes of geographical indications (GIs) by the applicants in rural specific regions primarily remain unexplored, although GIs can facilitate retention of long-term accumulated traditional knowledge in ethnic foods. This article sheds light on (1) analysis of phase-wise internal processes of GI registration and (2) examine relationships between the duration of application and perceived effects of GI registration in Northwestern Japan. GI in Japan is a comparatively a new system introduced in 2015. Methods: This article applied thematic analysis. The criteria for evaluating the gap of GI registration effects among the seven GI cases were descriptive and analytical with the information and data coming from official documents, questionnaire survey, and personal communications with key informants of the GI applicants. Qualitative information with quantitative data was also applied to visualize the complex internal GI registration processes and their effects. Results: This article identified that GI applicants who took a relatively long time to complete GI registration tended to perceive the relatively weak registration effects. Alternatively, GI applicants who took a relatively short time to complete their GI registration tended to evaluate the effects of the registration highly. For the preparation process, the most difficult phases for applications were drafting the application documents and long-term communication with a GI regulator. Conclusion: This article provides insights on the methods to uncover the complex, both ex-ante and expost, GI registration processes to reorganize current members of GI applicants and to fill the gaps and enhance the effects of GI registration.

This paper presents a local implementation of the Retrieval-Augmented Generation (RAG) model for private group messaging systems that require scalability, long-term conversation history storage, and low retrieval latency. The paper explains the integration of the RAG model with query decomposition, multi-hop retrieval, and hybrid reranking to support context-aware reasoning over chat history without external services. Performance evaluation showed retrieval latency below 100 ms for 100 concurrent queries when processing 50,000 stored messages on multiple hardware platforms. The study compared answers from RAG-based models and a non-RAG model using 150 human-verified QA pairs. All RAG-based models produced factual answers with a faithfulness score of at least 0.8, and the Qwen-3 (8B) model achieved the highest semantic accuracy of 0.86. The results indicate a significant difference in scores between the non-RAG and RAG models, demonstrating that retrieval and indexing capabilities are necessary for highly accurate and faithful responses.

Mobile Government is one of the new and important developments in e-government. The promise of m-government to provide greater access to government information is progressing in many developed and developing countries. Current Mobile Government does not exploit the full potential of available technology. Research shows there is a lack of a perfect killer application [2][1]. Firstly, this paper begins with an overview of e-gov and m-gov. In this context, this paper aims to present a framework to find killer applications in M-government to which researchers must pay attention and develop suitable way to perform according to different countries' contexts.

The Industrial Control System (ICS) is an environment composed of control systems and field devices used to automate industrial processes. With the technological development of the 4th Industrial Revolution, as the connection between the ICS and the external network expands, the risk of exposing cyber threats to the facilities is increasing. Accordingly, cyber-attack response exercises are being implemented around the world using cyber-attack scenarios. A lot of studies are being conducted on evaluating cyber-attack response exercises as well. However, most studies focus only on evaluating the performance of response activities rather than evaluating the cyber-attack scenarios used for these exercises. Our research presents a quantitative evaluation framework to evaluate the quality of cyber-attack scenarios used in cyber-attack response exercises. Our proposed framework consists of a total of 2 stages and determines whether to use the cyber-attack scenario for cyberattack response exercise.

고객 경험 디자인은 고객이 제품 또는 서비스를 구매하고 사용할 때 브랜드나 기업의 터치포인트에서 긍정적인 경험을 최대화하는 것을 목표로 한다. 긍정적인 경험을 전달할 수 있는 제품이나 서비스를 디자인하기 위해서는 고객의 니즈와 고객 경험의 변화하는 특성을 해결하는 데 도움이 될 수 있는 다른 관점과 접근 방법이 필요하다. 디자인 프로세스 는 다양한 학문 분야에 걸쳐 혁신을 주도하고 브랜드나 기업의 지속적인 성장에 필수적인 접근 방법 중 하나이다. 따라서, 본 연구에서는 고객과의 상호 작용과 고객 경험을 개선하는 것을 목표로 하는 서큘러 고객 경험 디자인 프로세스(The Circular CX Design Process)를 제안한다. 본 연구에서 제안한 프레임워크는 고객 여정을 중심으로 다양한 터치포인트 를 분석하고 개선함으로써 지속적으로 고객에게 긍정적인 경험을 제공하는데 그 목적이 있다.

Current research and practice have increasingly sought to focus on approaches to facilitate innovation due to its expanding growth of the experience economy over the past decade. Normally, customer experience design aims to maximize the effectiveness of positive experience with a brand or company's touchpoints when customers make a purchase as well as use a certain product or service. In order to design a product or service that can provide varying levels of experiences, there needs more beneficial design process strategies to help resolve the changing nature of customer experience. Design process is one such approach that provides designers as well as researchers across various academic disciplines as an integral source of creativity, driving innovation and growth. In this regard, this study aims to look specifically at the nature of customer experience and attempts to understand what extent that design process in both practice and theoretical perspective can assist in customer experience improvement. As a further consequence, this study will focus on implementing the idea of circular design into the customer experience process, namely the Circular CX Design Process―which is expected to increase effectiveness and efficiency in addressing customer needs and demands as well as delivering positive experiences at various touchpoints along the customer journey.

Recently, as evaluation of information security (IS) management become more diverse and complicated, the contents and procedure of the evidence to prepare for actual assessment are rapidly increasing. As a result, the actual assessment is a burden for both evaluation agencies and institutions receiving assessments. However, most of them reflect the evaluation system used by foreign government agencies, standard organizations, and commercial companies. It is necessary to consider the evaluation system suitable for the domestic environment instead of reflecting the overseas evaluation system as it is. The purpose of this study is as follows. First, we will present the problems of the existing information security assessment system and the improvement direction of the information security assessment system through analysis of existing information security assessment system. Second, it analyzes the technical guidance for information security testing and assessment and the evaluation of information security management in the Special Publication 800-115 'Technical Guide to Information Security Testing and Assessment' of the National Institute of Standards and Technology (NIST). Third, we will build a framework to implement the evidence collection system and present a system implementation method for the ‘6. Information System Security’ of ‘information security management actual condition evaluation index’. The implications of the framework development through this study are as follows. It can be expected that the security status of the enterprises will be improved by constructing the evidence collection system that can collect the collected evidence from the existing situation assessment. In addition, it is possible to systematically assess the actual status of information security through the establishment of the evidence collection system and to improve the efficiency of the evaluation. Therefore, the management system for evaluating the actual situation can reduce the work burden and improve the efficiency of evaluation.

본 연구는 공공기관 기록관리 평가 정책의 타당성과 실효성을 진단하고, 이를 위한 메타평가 설계 방안을 제안하였다. OECD DAC의 6대 평가 기준을 바탕으로 ‘구조적 설계’, ‘절차적 타당성’, ‘성과 환류 가능성’의 세 범주로 분석틀과 체크리스트를 구성하였으며, 2023년과 2024년 기관평가 결과를 적용해 유형별 한계를 비교·진단하였다. 분석 결과, 평가지표의 일관성 부족, 환류 체계 미비 등 구조적 문제가 확인되었으며, 이는 메타평가가 제도 개선을 위한 유효한 도구로 기능할 수 있음을 시사한다. 본 연구는 메타평가 개념을 기록관리 정책에 구체적으로 적용한 선도적 시도로서 의의를 갖는다.

This study aims to evaluate the validity and effectiveness of the public records management evaluation policy and proposes a meta- evaluation framework to assess it from a higher-level perspective. Based on the six OECD DAC evaluation criteria, the study constructs an analytical model and checklist categorized into three dimensions: structural design, procedural validity, and institutional feedback. Applying this framework to the 2023 and 2024 institutional evaluation results, the study identifies structural limitations such as inconsistent indicators and the absence of effective feedback mechanisms. These findings demonstrate that meta-evaluation can serve as a practical tool for improving evaluation systems. This research represents a pioneering effort to apply meta-evaluation to public records management policy in Korea.

디지털 기술의 확산과 활용이 증가함에 따라 민감한 데이터 보호와 시스템 무결성 강화를 위한 안티탬퍼링 기술의 중요성이 더욱 부각되고 있다. 본 연구는 하드웨어 및 소프트웨어 기반 안티탬퍼링 기술을 패키징, 보드 및 칩, 소프트웨어로 분류하여 각 기술의 특성과 평가 방법을 체계적으로 분석하였다. 또한, 반응 시간, 탐지 정확도, 물리적 내구성, 환경 적응성 등 주요 평가 지표를 기반으로 통합적인 평가 체계를 제안하였다. 본 연구에서 제시한 체계는 안티탬퍼링 기술의 효과를 객관적이고 표준화된 방식으로 검증할 수 있는 기준을 마련하며, 방위 산업, 금융, IoT 기기 등 다양한 첨단 응용 분야에서의 보안 기술 설계 및 검증을 지원하여 진화하는 사이버 위협에 대한 대응력을 강화할 것으로 기대된다.

With the rapid proliferation of digital technologies, the importance of anti-tampering measures for safeguarding sensitive data and ensuring system integrity has become paramount. This study conducts a systematic analysis of hardware- and software-based anti-tampering technologies, categorizing them into four key domains: packaging, board and chip, code, and software. The characteristics and evaluation methodologies of each category are comprehensively analyzed to provide a holistic understanding of the field. To facilitate a standardized assessment, this paper comprehensively analyzes key metrics such as response time, detection accuracy, physical durability, and environmental adaptability in an integrated manner. Our analysis serves as a structured and objective methodology for evaluating the effectiveness of anti-tampering technologies, enabling a consistent evaluation for comparison and analysis. Furthermore, our research is intended to guide the design and validation of security technologies across a wide range of applications, including defense systems, financial services, and IoT devices. This study aims to enhance the robustness and reliability of security solutions, paving the way for more resilient protection against increasingly sophisticated tampering techniques.

본 논문은 데스크탑에서 MCU의 자원·전력·지연 제약을 등가(iso-resource)로 근사 재현해 온디바이스 성능을 추정하고 보고하는 Desktop-to-Device(D2D) 프레임워크를 제안한다. 모델을 연산자 단위로 요약한 op_summary.json과 MCU 특성을 담은 mcu_profile.json을 결합하여 지연(P50/P95/P99), SRAM/Flash, 추론당 에너지, EPcD, deadline miss(%)를 표준 형식으로 산출한다. 전처리와 윈도 갱신을 포함한 E2E 경로를 평가하며, 마이크로벤치마크로 LUT 및 테일 계수(α, β)를 교정해 예측-실측 오차를 제시한다. 두 IDS 태스크에서 데스크탑 지표/나이브 MACs 스케일링이 MCU 지연을 과소추정하거나 비교를 왜곡한 반면, D2D는 L/M/H 프로파일 간 지연과 자원 및 에너지 차이를 보드 없이도 일관되게 제공하고 저사양 환경의 비선형 지연 증가와 자원 적합성을 현실적으로 반영하였다.

This paper proposes a Desktop-to-Device(D2D) framework that estimates and reports on-device performance by approximately reproducing MCU resource, power, and latency constraints on a desktop in an iso-resource manner. By combining op_summary.json, which summarizes a trained model at the operator level, with mcu_profile.json, which captures MCU characteristics, the framework outputs standardized metrics including latency (P50/P95/P99), SRAM/Flash usage, energy per inference, EPcD, and the deadline miss rate (%). The framework evaluates the end-to-end (E2E) path, including preprocessing and window updates, and uses a microbenchmark to calibrate the operator lookup table (LUT) and tail-latency coefficients (α, β), reporting the resulting prediction-measurement error. Across two IDS tasks, desktop-only measurements and naive MACs scaling systematically underestimated MCU latency or distorted relative comparisons, whereas D2D consistently provided latency, resource, and energy differences across L/M/H profiles without requiring a physical board, realistically capturing non-linear latency increases and memory-fit constraints in low-end settings.

Purpose: This study was conducted to develop an evaluation framework for QI activity in medical institutions and to analyze QI activity cases by applying the developed evaluation framework. Method: A four-phase process was employed to develop the evaluation framework, and a descriptive survey was used for the QI case study. Data were collected in April, 2010 by examining 157 QI activity cases presented at conferences and published in Journal of Korean Society of Quality Assurance in Health Care over the past three years. Developed QI activity evaluation instruments were used for data collection. Data were analyzed using the SPSS 18.0 for Windows program. Result: A QI Activity Evaluation Framework was developed. This framework consisted of 45 items. The department with the highest level of QI participation was the nursing department. The most frequent QI activity theme was patient safety. QI activity levels in Korean medical institutions are relatively equalized without significant differences according to institution characteristics. Conclusions: From the quality aspect of QI activity, more systematic and scientific approaches are required to upgrade QI activity. This study could provide methodological guidelines for QI activity and be useful in setting goals and directions for QI activity in medical institutions in Korea.

건설공사 사후평가는 공공건설공사의 효율성을 평가하고, 사업결과에 대한 적정성을 검증하여 추후 유사한 건설공사에 활용 가능한 정보를 제공하기 위해 도입되었으나, 인식 부족 등에 의해 활발한 제도 이행은 미흡한 실정이다. 이에 본 연구는 건설공사 사후평가 수행결과를 활용하여 추후 유사한 프로젝트 수행 시 참고자료로 활용 가능한 기본적인 틀을 제공하는 것을 목적으로, 건설공사 사후평가 수행결과의 효과적인 분석방법을 제시하고 이를 토대로 시범적인 분석을 수행하였다. 분석은 사후평가 결과가 건설CALS에 입력된 국도, 고속도로, 철도, 항만 등 293건의 건설공사를 대상으로 하였다. 분석결과, 철도 사업의 공사비 증감율이 고속도로, 국도, 항만공사 보다 높게 나타났다. 그리고 고속도로의 공사기간 증감율은 철도, 국도, 항만보다 큰 것으로 나타났다. 본 연구를 통해 제시된 사후평가 분석 프레임워크는 추후 프로젝트에 유용한 정보를 제공하고 벤치마킹 도구로써 유용하며, 지속적인 건설공사 사후평가 수행을 통해 다양한 분석이 가능하다는 점에서 큰 의미가 있다.

A construction post evaluation is introduced to provide information that can be used for similar construction projects in the future through the evaluation of the efficiency of a public construction project and the verification of the appropriateness of the project results. However, it is not widely used due to a lack of awareness of the method. For this reason, the aim of this study is to provide a basic frame that can be utilized as a reference for a similar project in the future based on the construction post evaluation. More specifically, a more effective analysis method for the construction post evaluation is proposed and then an exemplary analysis is performed based on the method. The paper analyzes 293 cases stored in construction CALS. Construction cost growth rates of railway projects are tend to be greater than other project types like highway, expressway, and harbour. Then, construction schedule grwoth rates of expressway are greater than others. This is expected to provide useful information for a future project, and serve as a basic frame. Moreover, it is also meaningful that a project can be analyzed from different perspectives through the construction post evaluation.

본 연구는 사이버 복원력을 정량적으로 평가하기 위한 지표 선정 프레임워크를 제시한다. 정보보호 및 개인정보보호 관리체계 인증기관을 기반으로 다양한 서비스 유형의 특성을 분석하고 복원력 측정을 위한 지표 후보군을 도출한다. 선정된 후보 지표는 본 연구에서 정의한 객관성, 재현성, 확장성, 실용성, 복원력 반영성의 다섯가지 핵심 특성을 기준으로 타당성을 검증한다. 또한 지표간 상호 배타성(ME)과 전체 포괄성(CE) 원칙을 보완 기준으로 적용하여 중복성을 제거하고 평가 체계의 범용성을 확보하였다. 최종적으로 12개의 정량적 평가 지표를 선정하였으며 이 중 웹서비스와 같은 트랜잭션 기반 시스템에서의 초당 처리 건수를 나타내는 TPS 지표를 중심으로, TCP SYN flooding 공격 상황에서의 지표의 변화와 해석방법에 대한 실증적 분석을 수행하였다. 제안한 지표선정 프레임워크와 정량적 평가지표는 사이버 복원력을 객관적으로 측정할 수 있는 평가 체계 수립에 기여하며, 지속적으로 변화하는 사이버 위협에 효과적인 대응과 개선을 위한 기준을 제공한다.

This study proposes a framework for selecting quantitative metrics to evaluate cyber resilience. Based on an analysis of various service types certified under information security and personal data protection management systems (ISMS-P), a pool of candidate resilience metrics was derived. The validity of the selected metrics was evaluated according to five key criteria defined in this study: objectivity, reproducibility, scalability, practicality, and resilience representation. In addition, the principles of mutual exclusivity (ME) and collective exhaustiveness (CE) were applied as supplementary criteria to eliminate redundancy and ensure the general applicability of the evaluation framework. As a result, 12 quantitative resilience metrics were selected. Among them, a detailed empirical analysis was conducted on the Transactions Per Second (TPS) metric, which measures transaction throughput in systems such as web services, focusing on its variation and interpretation under TCP SYN flooding attack scenarios. The proposed metric selection framework establishes a standardized evaluation framework for objectively measuring cyber resilience, thereby providing a basis for effectively responding to and enhancing resilience against continuously evolving cyber threats.

DeFi 생태계 확산은 스마트 컨트랙트 접근 제어 취약점을 통한 대규모 금전 손실을 가속하고 있다. 기존 정적·동적 분석 도구는 코드 단위 취약점에는 효과적이나, 프로토콜 구조와 권한 흐름에서 발생하는 논리적 취약점 탐지에는 한계가 있다. 본 논문은 스마트 컨트랙트 IR(Intermediate Representation)을 기반으로 연산 흐름, 상태 변화, 권한 검증 패턴을 통합 분석하는 지능형 정적 분석 프레임워크를 제안한다. 제안 기법은 권한·호출·상태 변경 패턴을 정량적으로 집계·정규화하고, IR 기반으로 통계 임곗값을 자동 조정하는 Semantics 기반 지능형 분석 방식을 통해 취약점을 판정한다. 2020-2024년 주요 해킹 사례 20건을 대상으로 한 평가에서 95%의 탐지율과 1.72초의 평균 지연 시간을 기록하며 기존 도구 대비 정확성과 효율성을 입증하였다. 본 연구는 규칙 기반 분석의 한계를 보완하고, 프로토콜 구조적 특성을 반영함으로써 DeFi 스마트 컨트랙트 접근 제어 취약점 검증의 정밀도를 실질적으로 향상시킨다.

The growth of the DeFi ecosystem has intensified financial losses caused by access-control vulnerabilities in smart contracts. Existing static and dynamic tools detect code-level issues but struggle to capture logical flaws tied to protocol structures and authorization flows. This paper introduces an intelligent static analysis framework that uses an IR to integrate operation flow, state transitions, and authorization-verification patterns. By aggregating and normalizing permission, call, and state-change patterns and applying a semantics-based method with automatically adjusted statistical thresholds, the framework identifies access-control weaknesses more effectively. Evaluation on 20 major hacking incidents from 2020-2024 shows a 95% detection rate and 1.72-second average latency, outperforming existing tools and substantially improving the precision of DeFi access-control vulnerability analysis.