Cloud-native databases are increasingly adopting disaggregated architectures to overcome limitations of traditional shared-nothing designs, enabling elastic scalability and cost efficiency. This shift allows compute and storage to scale independently, paving the way for innovations like pushdown processing and specialized hardware integration. As research advances, disaggregation promises to redefine database resilience, performance, and adaptability for modern applications.