Comparison between ApsaraDB for MongoDB and self-managed databases - ApsaraDB for MongoDB

Compared with self-managed databases, ApsaraDB for MongoDB has outstanding advantages in terms of service availability, data reliability, security, and O&M cost. ApsaraDB for MongoDB helps you quickly start your business and reduce O&M costs.

Item	ApsaraDB for MongoDB	Self-managed database
Service availability	High availability.	You must make sure of security and build primary/secondary replication and RAID.
	High availability disaster recovery in one or three zones in the same region.	Zone-disaster recovery requires self-deployment and maintenance. The dual-zone conditions are difficult to implement, and database-level availability is not guaranteed.
	Allows you to create geo-disaster recovery instances.	You must use a third-party tool to create disaster recovery instances.
Data reliability	High reliability. The recovery point objective (RPO) of single- and three-zone replica set instances is 0.	You must ensure security and build primary/secondary replication and RAID.
System security	Pre-protection: DDoS attack protection, automatic repair of various database security vulnerabilities, whitelist-based access control, and VPC isolation.	Pre-protection: requires additional security hardware or software to fix security vulnerabilities at high costs.
	In-process protection: SSL encryption and Transparent data encryption.	In-process protection: You must build SSL encryption and TDE systems.
	The SQL Audit feature is provided. For more information, see View audit logs.	Event auditing: You must purchase an additional auditing system.
Backup and restoration	Complete kernel enables both physical backup and logical backup to be performed in manual backup and increases backup efficiency three times. Single-database recovery.	The open-source version supports only logical backup and is inefficient. Single-database recovery is unavailable and recovery efficiency is low. You must ensure the accuracy of data recovery in the distributed architecture.
System hosting	No hosting fee.	You must pay additional server hosting fees. The more complex the architecture, the more servers you need to host and the higher the hosting fees.
O&M cost	No additional O&M costs.	Requires dedicated maintenance personnel and high O&M costs.
O&M cost	The following CloudDBA intelligent diagnosis features are supported: Performance trends Real-time performance Instance sessions Slow query logs Storage analysis	No performance diagnosis is available and it is very complex to troubleshoot slow queries.
Deployment and scaling	Instantly activated and supports auto scaling.	Requires hardware procurement, hosting of data centers, and machine deployment, which takes rather a long period. You must maintain the node relationship when you add new nodes.
Kernel optimization	Optimizes oplog synchronization performance and short-lived connection performance for general scenarios. Ensures data consistency with technologies such as logical snapshots. This reduces costs by 50% in long-time data import scenarios.	The open source version does not provide optimization and cannot be used in some scenarios.