A maturity model for AI-empowered cloud-native databases: from the perspective of resource management

Feng, Xiaoyue; Guo, Chaopeng; Jiao, Tianzhe; Song, Jie

doi:10.1186/s13677-022-00318-1

Journal of Cloud Computing

Advances, Systems and Applications

Table 7 Interleaving of CNDB and AI dimensions in our study

From: A maturity model for AI-empowered cloud-native databases: from the perspective of resource management

AI dimension	CNDB dimension
AI dimension	Multi-tenant	Compute and storage disaggregation	Cross-Az/Region	Near-data processing	Logs as the database	Distributed and shared memory
Smart resource prediction	Databases provide corresponding resources according to the needs of each tenant, and AI technologies can be applied to predict the resources that each tenant requires.	Databases predict the demand of computing and storage resources under different application backgrounds.	The resource management system can predict the demand of resources in different regions.	Databases predict the computing resources required for the operator push-down in the storage layer.	Databases apply AI technologies to predict the storage resources required to store logs.	Databases predict the shared memory of each machine according to the historical resource usage.
Smart resource scheduling	Databases can apply AI technologies to dynamically schedule resources according to the resource demand of tenants in the use process.	Databases can dynamically adjust the decoupling degree of storage and computing resources according to different application backgrounds.	Databases can dynamically schedule resources based on resource usage in different regions.	Databases can dynamically adjust the computing resources in the storage layer due to the operator push-down	Databases can schedule storage resources in real-time based on the resources required to store logs.	Databases can schedule the memory resources in real-time based on the results of resource prediction to maintain the shared memory pool.
Smart resource control	Databases have a complete resource control system to control the resource usage and consumption in all parts of the database. It can ensure that resources will not be over-utilized or under-utilized.
Smart resource scaling	Database systems can decide when to add new resources and automatically bring resources into the scope of resource management. It realizes the automatic scaling function of resource containers.

Back to article page