Constructing energy-efficient database systems to reduce economic costs and environmental impact has been studied for 10 years. With the emergence of the big data age, along with the data-centric and data-intensive computing trend, the great amount of energy consumed by database systems has become a major concern in a society that pursues Green IT. However, to the best of our knowledge, despite the importance of this matter in Green IT, there have been few comprehensive or systematic studies conducted in this field. Therefore, the objective of this article is to present a literature survey with breadth and depth on existing energy management techniques for database systems. The existing literature is organized hierarchically with two major branches focusing separately on energy consumption models and energy-saving techniques. Under each branch, we first introduce some basic knowledge, then we classify, discuss, and compare existing research according to their core ideas, basic approaches, and main characteristics. Finally, based on these observations through our study, we identify multiple open issues and challenges, and provide insights for future research. It is our hope that our outcome of this work will help researchers develop more energy-efficient database systems.

With database management systems becoming complex, predicting the execution time of graph queries before they are executed is one of the challenges for query scheduling, workload management, resource allocation, and progress monitoring. Through the comparison of query performance prediction methods, existing research works have solved such problems in traditional SQL queries, but they cannot be directly applied in Cypher queries on the Neo4j database. Additionally, most query performance prediction methods focus on measuring the relationship between correlation coefficients and retrieval performance. Inspired by machine-learning methods and graph query optimization technologies, we used the RBF neural network as a prediction model to train and predict the execution time of Cypher queries. Meanwhile, the corresponding query pattern features, graph data features, and query plan features were fused together and then used to train our prediction models. Furthermore, we also deployed a monitor node and designed a Cypher query benchmark for the database clusters to obtain the query plan information and native data store. The experimental results of four benchmarks showed that the average mean relative error of the RBF model reached 16.5% in the Northwind dataset, 12% in the FIFA2021 dataset, and 16.25% in the CORD-19 dataset. This experiment proves the effectiveness of our proposed approach on three real-world datasets.

Given a multidimensional dataset, the skyline query is to find all objects that are not dominated by any other object in the dataset. This query represents a multi-objective optimization approach that is used extensively in various areas such as data analysis, decision support, and intelligent recommendation systems. However, even the most effective skyline query algorithms in the literature often show some limitations in dealing with diverse distribution data. By further analyzing the relationship between index structure and dominance tests, we propose the Quad*-Tree, a variant of the Quad-Tree that allows for configurable fan-out and addresses the issue of low node utilization in high-dimensional spaces. Based on Quad*-Tree, we introduce the concept of Node Dominance Groups (NDG). Based on NDG, we develop U-INDG (Upward Identification of Node Dominance Groups), a suite of novel and optimal skyline query approaches that significantly reduce the cost of dominance tests. Subsequently, we extend our approach to other index structures such as the sorted R-Tree, ZB-Tree, and Quad-Tree. Furthermore, U-INDG is simple to implement and can be efficiently scaled to various systems. Extensive experiments show that U-INDG outperforms the most efficient index-based skyline algorithms currently available. It significantly reduces query time across various types of datasets, with speedups ranging from approximately 2.4\(\times \) to 7.8\(\times \) compared to state-of-the-art algorithms, while ensuring stable performance.