Some businesses operate in multiple regions. You can deploy MongoDB instances in multiple regions according to the business distribution to support multi-region business operations. A typical deployment uses MongoDB in one of the two regions as the main business data source where the applications write data. MongoDB in the other region mainly reads and synchronizes data from MongoDB in the main business region. This solution installs and deploys the MongoShake service, an open-source project developed and maintained by Alibaba on the ECS of the main business region. MongoShake supports the large-scale, cross-regional, e-commerce business of Alibaba Group to achieve one-way data synchronization from MongoDB in the main region to MongoDB in the other region. The synchronization latency depends on the instance load and network latency. The ideal latency control for the generated environment is milliseconds.

The following architecture diagram demonstrates how to deploy MongoDB instances in two business regions.

1. Install MongoDB instances in the specified regions. Then, install the ECS instance in the main business region where the source MongoDB is. MongoShake will be installed on this ECS instance later. In this solution, we use Terraform to build a MongoDB replica set and an ECS instance.
2. Configure the related security settings, such as accounts, security groups, and public network addresses on MongoDB and ECS .
3. Install and configure MongoShake on ECS.
4. Activate the MongoShake service and run sample applications to read and write from MongoDB. Observe and check the data duplication process.
5. Refer to the Terraform Template for Building Resources to configure Terraform.
6. Refer to the Python Scripts for Monitoring Data Synchronization to write the Python scripts for monitoring the data synchronization of MongoDB.
7. Refer to the Deployment Guide to complete the deployment process step by step.