Performance tuning

IBM® Sterling Order Management System provides seamless performance that is based on the foundational design principles of horizontal scalability, asynchronous workload management, and modular services.

Sterling Order Management System is deployed on the Kubernetes platform that provides a robust mechanism of workload management that enables the automatic adjustments to handle increasing API and user traffic.

You can also purchase more processing power for industry events such as the Back to school season, the retail Black Friday, or the Cyber 5 holiday peak.

The workload is divided into two categories:
  • Synchronous API response patterns that require immediate response.
  • Asynchronous background workload processes called agents and integration servers that process large workloads of data processing and communication by reading or retrieving data from different systems.

    This unique design is available in Sterling Order Management System where agents are automatically scaled up and scaled down and the workload is distributed across multiple JVMs to achieve infinite scale. This design is also anchored on the Kubernetes platform for the extended benefit.

Akamai based edge acceleration is supported for some of the modular capabilities that use the Akamai’s CDN Intelligent Edge CDN platform with over 3,65,000 servers across 135 countries. It provides optimized routes and API caching that further improves the user experience.

Sterling Order Management System provides the option to achieve more improvements and to achieve the best performance objectives. For more information about performance profiles, see Configuring servers on the next-generation platform.

Tuning and performance activities

As part of tuning and testing your Sterling Order Management System performance, complete the following activities:
  1. Configure your environment for performance testing. Most of your initial testing before your service is started, such as single user testing, is typically completed within your quality assurance environment. Complete more robust load testing within your preproduction environment. After your site is started, use the quality assurance environment if you need to complete more performance testing.
  2. Review profiling and caching results.
  3. Run test scripts and monitor system responses. The load test scripts implement the main usage scenarios that are defined in the test plan. If you need to adjust the workload mix that is used for the scripts, adjust the number of virtual users that are allocated per script. Adjusting the workload mix is an iterative process. The goal is to simulate the correct number of requests and visitors with the correct request distribution corresponding to the test plan.
  4. Identify and tune any identified problem areas. To help identify problem areas, review log files and use debugging and test tools. For example, to help with browser-side debugging, you can use the Firebug browser plug-in. Performance testing tools such as load generators are not included with your Sterling Order Management System service. These performance testing tools must be installed and run outside of the Sterling Order Management System architecture.
  5. Rerun load scripts to unmask the underlying problems.
  6. Document the test results.
  7. Tune any identified problem areas again. Then, repeat the test and tuning process.

Performance test flow

performance test flow: single user, single system concurrency, incremental scale, stability

For each task flow, follow this performance test flow.

Simulate and measure the performance of:
  1. A single user;
  2. Concurrent, multiple users on single system;
  3. Incremental scaling (multiple users on multiple systems);
  4. Stability testing (multiple users, on multiple systems, over time).