Data backup is a core technology for improving system resilience to system failures. Data backup in enterprise systems is required to minimize the impacts on business processing, which can be categorized into two factors: system slowdown and downtime. To eliminate system slowdown, asynchronous data copy (ADC) technology is prevalent, which copies data asynchronously with original data updates. However, the ADC can collapse backup data when applied to enterprise systems with multiple resources. Then, the demonstration system employed consistency group technology, which makes the order of data updates the same between the original and backup data. In addition, we developed a container platform operator to unravel the complicated correspondence between storage volumes and applications. The operator automates the configuration of the ADC with the setting of consistency groups. We integrated the storage and container technologies into the demonstration system, which can eliminate both system slowdown and downtime. The demonstration video can be accessed at https://photos.app.goo.gl/CkocfRasTUxuzzgr5.

Demand for data transformation has increased with the rapid growth of data. To ensure improvements in production line efficiency, Internet of Things (IoT) data analytics demand data transformations that include changing the format, structure, or values of the data stored in a data lake. Data transformations are generally not reused due to the complexity of data transformation files and the lack of knowledge on previous data transformation flows developed by other Extract Transform Load (ETL) developers who develop data transformations. For a naive developer, it is time consuming and difficult to find relevant existing data transformations that can be modified as per new requirements. To solve this problem, we developed a knowledge graph-based data transformation recommendation system featuring a data similarity component that helps to provide explainable results. This system, can improve the mean average precision 26% and the mean average recall by 24% while reducing the mean average root mean squared error by 69% which implies that it can significantly increase the effectiveness of the of recommendations. With the help of this recommendation engine, data transformation tasks can be done by naive developers with little knowledge on existing data transformation flows and continuous improvement projects can be speeded up.

Overfitting reduces the generalizability of deep neural networks (DNNs). Overfitting is generally detected by comparing the accuracies and losses of training and validation data; however, the detection method requires vast amounts of training data and is not always effective for forthcoming data due to the heterogeneity between training and forthcoming data. The dropout technique has been employed to prevent DNNs from overfitting, where the neurons in DNNs are invalidated randomly during their training. It has been hypothesized that this technique prevents DNNs from overfitting by restraining the co-adaptions among neurons. This hypothesis implies that overfitting of a DNN is a result of the co-adaptions among neurons and can be detected by investigating the inner representation of DNNs. Thus, we propose a method to detect overfitting of DNNs using no training and test data. The proposed method measures the degree of co-adaptions among neurons using persistent homology (PH). The proposed PH-based overfitting measure (PHOM) method constructs clique complexes on DNNs using the trained parameters of DNNs, and the one-dimensional PH investigates the co-adaptions among neurons. Thus, PHOM requires no training and test data to measure overfitting. We applied PHOM to convolutional neural networks trained for the classification problems of the CIFAR-10, SVHN, and Tiny ImageNet data sets. The experimental results demonstrate that PHOM reveals the degree of overfitting of DNNs to the training data, which suggests that PHOM enables us to filter overfitted DNNs without requiring the training and test data.

The in-memory system is promising for improving the performance of column-oriented database management systems (DBMSs). However, in comparison with NAND-flash-based solid-state drives (SSDs), dynamic random access memories (DRAMs) are around ten times more expensive and tens to thousands of times more energy inefficient. To overcome this drawback, we developed a column-oriented DBMS and a field-programmable-gate-array-based acceleration engine. We integrated them in FCAccel, our prototype system. The acceleration engine accelerates data extraction from SSDs for SQL processing. We compared the performance of FCAccel with that of MonetDB, Impala and PostgreSQL. Performance was evaluated under the conditions that MonetDB and Impala stored all data in DRAMs and FCAccel stored all data in SSDs. With regard to data extraction, the performance of FCAccel ranged from 0.77 to 1.79 times in comparison with that of MonetDB and from 4.10 to 23.4 times in comparison with that of Impala. These experimental results imply that the acceleration engine can remove the necessity to store all data in DRAMs.