X

The Oracle NoSQL Database Blog covers all things Oracle NoSQL Database. On-Prem, Cloud and more.

  • January 27, 2020

Multi-Region Data Replication

Tim Goh
Principal Product Manager

In today’s competitive global market, businesses are facing the challenge of delivering faster and better services to their customers. Enterprises need to develop new innovative applications and make them available globally in the shortest time possible. The complexity of deploying, operating, and maintaining these applications in multiple regions to serve the global customers can be daunting. Users expect to complete their online activities with smooth and fast user experience at any time and from anywhere. To meet such expectations, enterprises need to host applications and data at distributed regions closest to the users. Data needs to be actively replicated across these regions with predictable low latency.

Oracle NoSQL Database is designed for today’s most demanding workloads with high volume, high velocity, and high variety. It meets the requirements of applications for  predictable low latency, multiple data models, rapid development, elastic capacity scaling, ease of operations, and management. Data and applications can reside anywhere and interoperate with a single application interface. The built-in multi-region data replication feature enables application data written in any region to be replicated transparently across multiple regions right away.

For example, a mobile application company would like to serve their customers in three different regions: Frankfurt, London, and Dublin. When a user updates his user profile via a smartphone in Frankfurt, the new data is replicated in London and Dublin immediately. Similar replication will be performed when an update happens in Dublin or London.  If updates happen at all regions, Oracle NoSQL multi-region architecture handles the complexity of conflicts and replicates the data accordingly.

The mobile company can easily set up the Oracle NoSQL multi-region data replication to serve their customers in different regions. The diagram below illustrates the underlying multi-region architecture. To enable multi-region data replication, NoSQL multi-region tables are created in Frankfurt, London, and Dublin clusters. These tables are read-anywhere and write-anywhere hosted in these regions.

One or a few Cross-Region Services can be deployed in each region. Each enables data streaming from remote regions to the local region. Each cluster in a region is fully autonomous and may have different topology, capacity, elastic expansion policy, operations, and maintenance.

Steps to Deploy Multi-Region Data Replication in a Region

Step 1: Deploy cross-region service in each region.

  • Create a directory for the Cross-Region (XRegion) Service in the local NoSQL cluster.
  • Create a JSON config file in the directory with the following parameters:

 

                       For example, a config file to stream data to London from Denmark and Dublin:

  • Start the XRegion Service in each region using the command:

For example, start XRegion Servce in London:

Step 2: Install Oracle NoSQL Database in each region using the standard steps and create regions.

Step 3: Create multi-region tables in each region.

Use CREATE TABLE to create a multi-region table. For example, use SQL command to create a multi-region table in London:

Step 4: Perform read and write operations.

After creating the multi-region table in each region, you are ready to perform read or write operations on the table using the existing data access APIs or DML statements. Try writing a few test records in London region and verify those records are automatically replicated in Dublin and Frankfurt regions.

Be the first to comment

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.