Extend your high availability database architecture to multicloud using OCI GoldenGate replication
This reference architecture describes a customer-inspired active-active database setup that provides a highly available architecture in a multicloud solution. It can be used in production operation, database migration or database maintenance.
Note:
With a multicloud solution, networking is a key determinant of overall system performance. It is the customer’s responsibility to ensure that Cloud-to-Cloud network (bandwidth and latency) is thoroughly tested to ensure the application performance meets the defined business requirements.Implementing an active-active setup provides the following benefits:
Optimal Utilization
Resources are used optimally instead of the database being in a passive state, thereby providing higher availability, distributing the workload, and improving performance for locally distributed users.
Cost and Flexibility
Multicloud solutions help customers choose cloud vendors based on regional availability and avoid any lock-in with a specific vendor, providing the best-in-class options.
Architecture
In this reference architecture, a copy of the production database is deployed in OCI US-East (Ashburn) and another copy is deployed in AWS US-East (N.Virginia) with OCI GoldenGate bidirectional data replication in an active-active setup between the two clouds.
Applications are deployed in AWS US-East (N.Virginia). Dedicated connectivity is provided by utilizing OCI Fast Connect via Megaport, an OCI FastConnect partner to cross connect OCI workloads to AWS.
A similar multicloud connection can be setup by any OCI FastConnect provider that serves the data center location, such as Equinix, AT&T, Lumen, NTT, Verizon, or any meetme floor with a telecom exchange provider.
The following diagram illustrates this reference architecture.
oci_multicloud_db_replication_goldengate-oracle.zip
In this setup, OCI GoldenGate is provisioned in OCI, and the OCI Autonomous Database and AWS Oracle database on EC2 are considered for bidirectional replication.
The service console for OCI GoldenGate provides a GUI to design and manage the processes (Extract and Replicat) and to analyze the performance metrics.
This architecture has the following components:
- FastConnect
Oracle Cloud Infrastructure FastConnect provides an easy way to create a dedicated, private connection between your data center and Oracle Cloud Infrastructure. FastConnect provides higher-bandwidth options and a more reliable networking experience when compared with internet-based connections.
- Autonomous Transaction
Processing
Oracle Autonomous Transaction Processing is a self-driving, self-securing, self-repairing database service that is optimized for transaction processing workloads. You do not need to configure or manage any hardware, or install any software. Oracle Cloud Infrastructure handles creating the database, as well as backing up, patching, upgrading, and tuning the database.
- Oracle Database on AWS - EC2
Oracle database services and products offer customers cost-optimized and high-performance versions of Oracle Database, the world's leading converged, multi-model database management system. Amazon RDS for Oracle is fully managed by Amazon Relational Database Service (RDS).
- OCI GoldenGate
Oracle Cloud Infrastructure (OCI) GoldenGate is a managed service providing a real-time data mesh platform, which uses replication to keep data highly available, and enables real-time analysis. Customers can design, execute, and monitor their data replication and stream data processing solutions without allocating or managing compute environments.
Recommendations
- VCN
When you create a VCN, determine the number of CIDR blocks required and the size of each block based on the number of resources that you plan to attach to subnets in the VCN. Use CIDR blocks that are within the standard private IP address space.
Select CIDR blocks that don't overlap with any other network (in Oracle Cloud Infrastructure, your on-premises data center, or another cloud provider) to which you intend to set up private connections.
After you create a VCN, you can change, add, and remove its CIDR blocks.
When you design the subnets, consider your traffic flow and security requirements. Attach all the resources within a specific tier or role to the same subnet, which can serve as a security boundary.
- Choice of interconnection location
This architecture requires one or more geographic locations for its components: the OCI region and associated OCI FastConnect edge node, the AWS region and associated AWS Direct Connect edge node. To achieve optimal end-to-end latency, we recommend selecting a metro with each of these architectural elements in close proximity.
- Application Design
When using active-active replication, the time zones must be same on both database systems so that timestamp-based conflict resolution and detection can operate.
- Data Uniqueness
When using active-active replication, set a unique range in sequences for easy identification to prevent conflicts. Conflict resolution procedures must be implemented on all systems in an active-active configuration. Conflicts should be identified immediately and handled with as much automation as possible. For more information, see the "GoldenGate conflict resolution capabilities" link in the Explore More section.
- Parallelism
It is recommended to set up parallel replication to speed up applying lag where possible.
Considerations
Consider the following points when deploying this reference architecture.
- Performance
Test and tune application queries in the database to have a uniform or better (Exadata optimizations) execution plan on both databases.
Test the replication apply speeds to ensure data is real-time on both sites. Application should be replication aware as much as possible.
For the customer use case in this reference architecture, the requirement was for data to be available on the secondary site within 15 minutes. With the help of OCI GoldenGate, data was consistently replicated in real time to the secondary site.
Network latency is vital for performance. Check and measure the network latency as part of application performance testing.
Network latency between applications and database hosted in different cloud data centers must be less than 10 ms. We recommend selecting a metro with the applications and database cloud data centers in close proximity to achieve optimal end-to-end performance.
For the customer use case in this reference architecture, network latency induced for the multicloud deployment was between 2-4 ms in OCI US-East.
- Impact Assessment
Replication design should consider impact on account of planned maintenance, data corruptions and site failures and have them tested and addressed as needed.
- Cost
OCI GoldenGate provides a web-based flexible and elastic service to move data in real-time to the Cloud.
- Conflicts and Data Handling
Application design is critical to ensure active-active replication can be set without conflicts. In this case, data for each source was generated based on a different sequence range. A good knowledge of the application is also necessary in handling conflicts.
In addition, technology should be easy to automate and handle any conflicts.