Vol. 51, No. 3,
Mar. 2024
Digital Library
A Reinforcement Learning based Adaptive Container Scheduling Back-off Scheme for Reducing Cold Starts in FaaS Platforms
Sungho Kang, Junyeol Yu, Euiseong Seo
http://doi.org/10.5626/JOK.2024.51.3.191
Function as a Service(FaaS) is a cloud computing service model that virtualizes computing resources and provides them in units of functions. As it enables flexible and easy service deployment, its use is rapidly growing in a cloudnative architecture. However, the initial execution of a function requested by a user in a FaaS platform involves several initialization steps, and this initialization overhead, that is, cold start, delays function execution. Our proposal is that when there is a request to execute the same function as the running function, waiting rather than immediately processing the request can reduce the occurrence of a cold start. In this paper, we propose a FaaS request waiting policy model based on reinforcement learning that pursues the best choice between sending and waiting for a function execution request. As a result of the comparison experiment with Openwhisk, the frequency of cold start reduced by up to 57% and the average execution time of the function reduced by up to 81%.
Homomorphic Encryption-Based Support Computation for Privacy-Preserving Association Analysis
Yunsoo Park, Lynin Sokhonn, Munkyu Lee
http://doi.org/10.5626/JOK.2024.51.3.203
Homomorphic encryption is a cryptographic scheme that enables computation on ciphertexts without decryption. Homomorphic encryption is attracting attention as a cryptographic technology that can solve the issue of user privacy invasion in machine learning and cloud services. A representative scheme of homomorphic encryption is the CKKS scheme. CKKS is an approximate homomorphic encryption scheme that supports real and complex number operations. In this paper, we propose a method to efficiently compute support among evaluation metrics of association analysis using CKKS scheme, and a method to compute supports in parallel using matrix multiplication for multiple itemsets. We implemented and evaluated the proposed method to compute supports using the HEaaN library. According to evaluation results, the support value calculated by the proposed method was almost identical to that calculated without encryption, confirming that the proposed method could effectively calculate the support value while protecting user data privacy.
Memory-Aware Eager Co-Scheduling for Multi-Tenant GPU Environments
Jeongjae Kim, Yunchae Choi, Hwansoo Han
http://doi.org/10.5626/JOK.2024.51.3.210
In a multi-tenant GPU environment, multiple applications are co-located on a single GPU to maximize utilization and throughput. However, co-location can lead to out-of-memory errors. Previous research addressed this problem by scheduling tasks that do not exceed the total GPU memory capacity. Our research introduces two novel methods that allow the co-location of additional tasks on a GPU while effectively preventing out-of-memory errors. Our approach involves immediate deallocation of unused memory within tasks, freeing up memory early on the GPU. This enables additional concurrent execution of multiple tasks on the GPU. Furthermore, by over-subscribing Unified Memory, tasks are scheduled to tolerate memory usage that exceeds the total GPU memory capacity. With our proposed schemes, it is feasible to reduce the execution time of multiple tasks compared to previous scheduling approaches and each scheme shows performance improvement of 7.3% and 1.9%, respectively, compared to prior research.
Abstractive Summarization Corpus Construction of National Assembly Minutes and Model Development
Younggyun Hahm, Yejee Kang, Seoyoon Park, Yongbin Jeong, Hyunbin Seo, Yiseul Lee, Hyejin Seo, Saetbyol Seo, Hansaem Kim
http://doi.org/10.5626/JOK.2024.51.3.218
The mainstream of summary research has been targeting documents, but recently, interest in meeting summary research has significantly increased. As part of the National Institute of Korean Language’s big data construction project, a study on the summary of the National Assembly minutes, which have not yet been studied in Korea, was conducted and a summarization dataset for the National Assembly minutes was constructed. Qualitative intrinsic human evaluation was conducted to verify the quality of the constructed dataset. In addition, by conducting quantitative and qualitative evaluations of datasets built through the generative summarization model, the evaluation of the National Assembly Minutes Summarization dataset and the research direction of future generative and minutes summaries were sought.
Cross-Project Defect Prediction for Ansible Projects
Sungu Lee, Sunjae Kwon, Duksan Ryu, Jongmoon Baik
http://doi.org/10.5626/JOK.2024.51.3.229
Infrastructure-as-Code (IaC) refers to the activities of automating overall management through code, such as creating and deploying infrastructure. Infrastructure-as-Code is used by many companies due to its efficiency, and many within-project defect prediction techniques have been proposed targeting Ansible, one of the IaC tools. Recently, a study on the applicability of Ansible"s cross-project defect prediction has been proposed. Therefore, Ansible’s cross-project defect prediction technique was used in this study, and its effectiveness was analyzed. Experimental results showed that the performance of the F1-based cross-project defect prediction was measured to be 0.3 to 0.5, and that it could be used as an alternative to the internal project defect prediction technique. It is therefore anticipated that this will be put to use in support of Ansible’s software quality assurance activities.
Graph Structure Learning: Reflecting Types of Relationships between Sensors in Multivariate Time Series Anomaly Detection
http://doi.org/10.5626/JOK.2024.51.3.236
Sensors are used to monitor systems in various fields, such as water treatment systems and smart factories. Anomalies in the system can be detected by analyzing multivariate time series consisting of sensor data. To efficiently detect anomalies, information about the relationships between sensors is required, but this information is generally difficult to obtain. To solve this problem, the previous work used sensor data to identify relationships between sensors, which were then represented using a graph structure. However, in this process, the graph structure only reflects the presence of relationships between sensors, not the types of relationships between sensors. In this pap er, we considered the types of relationships between sensors in graph structure learning and analyzed multivariate time series to detect anomalies in the system. Experiments show that improving detection accuracy in graph structure learning for multivariate time series anomaly detection involves taking into account the different kinds of relationships among sensors.
A Study on Reduction of False Alarms in Weapon System Software Static Test Using Natural Language Processing Model
Insub Lee, Hyoju Nam, Namhoon Jung, Kyutae Cho, Sungkyu Noh
http://doi.org/10.5626/JOK.2024.51.3.244
Recently, Securing software stability has become increasingly important as military systems have been upgraded. To this end, the Defense Acquisition Program Administration conducts reliability tests for weapon system software through static analysis tools. However, many false alarms occurred during the test process, resulting in a waste of time and resources. This paper aims to achieve a high positive/false positive classification rate by creating a dataset using the log of a static analysis tool and training a language model. Additionally, data processing methods appropriate for the static analysis features of weapon system software were investigated and analyzed during the research. As a result of the analysis, it was found that the CodeBert model pretrained in C/CPP and natural language using Optuna, a hyperparameter tuning tool, showed 4-5% higher performance based on the F1 score than the existing SoTA model. If the model presented in this research is mainly employed in software static testing, a significant number of false positives can be found.
Design and Implementation of a Tactical-datalink Unit for Interfacing and Forwarding(TUF) to Interlock and Forward MIDS JTRS
Sangtae Lee, Jongseo Kim, Taegwon Kim, Youngseung Kim, Seungbae Jee, Jaeyoung Cheon, MinGyu Jung
http://doi.org/10.5626/JOK.2024.51.3.252
The MIDS LVT terminal is changed to MIDS JTRS terminals with improved crypto modernization, enhanced throughput and frequency remapping functions due to encryption keys and bandwidth limitations. The Korean military is carrying out various performance improvement projects for the Link-16 network equipped with a MIDS JTRS terminal. In this paper, a tactical datalink unit for interfacing and forwarding(TUF) was designed and implemented to interlock and forward tactical information to the existing C2 host system (MCRC, KTMO, etc.) using MIDS JTRS terminal.
The TUF had improved scalability and maintainability through a server-client structure design that could remotely control and manage MIDS JTRS terminal interlocking and forwarding in consideration of the network operating environment. The TUF consisted of interlocking and forwarding, a basic C2 (command & control) host, and a monitoring tool. The TUF was verified through linkage with previously verified overseas tools. The TUF secured technical know-how for MIDS JTRS terminal integration by achieving the purpose of interlocking and forwarding through domestic technology development, away from the existing method of interlocking and forwarding through overseas tools. The existing weapon system without MIDS JTRS terminal linkage function could join the Link-16 network to enhance military operational operability and survivability.
Model Architecture Analysis and Extension for Improving RF-based Multi-Person Pose Estimation Performance
http://doi.org/10.5626/JOK.2024.51.3.262
An RF-based multi-person pose estimation system can estimate each human posture even when it is challenging to obtain clear visibility due to obstacles or lighting conditions. Traditionally, a cross-modal teacher-student learning approach has been employed. The approach utilizes pseudo-label data acquired by using images captured concurrently with RF signal collection as input for a pretrained image-based pose estimation model. In a previous research study, the research team applied cross-modal knowledge distillation to mimic the feature maps of image-based learning models and referred to it as "visual cues." This enhanced the performance of RF signal-based pose estimation. In this paper, performance is compared based on the ratio at which the learned visual cues are concatenated, and an analysis of the impact of segmentation mask learning and the use of multiframe inputs on multi-person pose estimation performance is presented. It is demonstrated that the best performance is achieved when visual cues and multiframe inputs are used in combination.
Correcting OCR Errors based on Error Patterns
Nara Kim, Yongseok Jo, Ho-Hyun Park
http://doi.org/10.5626/JOK.2024.51.3.271
The development of Optical Character Recognition (OCR) has made it possible to digitize analog documents. It shows very high recognition accuracy for standardized documents. However, OCR errors still occur frequently in complex documents. To resolve these issues, an OCR error correction procedure is required. The majority of OCR errors are repeated for the same characters. Accordingly, OCR error information has an important meaning in OCR error correction work. However, there are few studies utilizing OCR error information. In order to identify patterns, this study examines OCR error data. It then suggests an OCR mistake correction technique based on neural machine translation. Experiments were carried out using the English dataset from the ICDAR 2017/2019 Post-OCR text correction competition in order to validate the proposed method. The experimental results showed that the model using OCR error information demonstrated a higher improvement rate than the model without OCR error information. It also showed up to 8%P improved results compared to the existing state of the art.
A Trie-based IoT Data Indexing Scheme Utilizing Temporal Prefixes
Jooyoung Yoon, Sunbeom Kwon, Young-Kyoon Suh
http://doi.org/10.5626/JOK.2024.51.3.280
As Internet of Things technology becomes more widespread, various techniques have been proposed to efficiently retrieve large amounts of IoT sensor data. One of the state-of-the-art techniques, ST-Trie, has been shown to have inferior query processing performance compared to other composite indexing techniques for queries with a narrow time range of less than 1 day. To address this concern, this paper proposes TS-Trie, a trie-based indexing technique using temporal prefixes. TS-Trie uses 20-bit temporal information as a prefix to convert three-dimensional spatiotemporal information into 64-bit one-dimensional key values. It also improves the existing search method by building a double-linked list of nodes after 20 bits. By applying the proposed TS-Trie to three real-world IoT datasets, we measured the performance of processing range queries, k-NN, and Top-k queries. As a result, TS-Trie could on average shorten the query time by about 50%, 40%, and 60%, compared to the existing methods, on the three datasets, respectively. With a high compression ratio of 86%, we further confirmed TS-excellent Trie"s space efficiency and showed an average indexing speed that was around four times faster than previous techniques.
Search
Journal of KIISE
- ISSN : 2383-630X(Print)
- ISSN : 2383-6296(Electronic)
- KCI Accredited Journal
Editorial Office
- Tel. +82-2-588-9240
- Fax. +82-2-521-1352
- E-mail. chwoo@kiise.or.kr