Testing Bottlenecks and Service Virtualization

The constant pressures of faster delivery and the increasing dependence on third party applications has complicated the testing process. Day by day it is getting tougher to test today’s composite applications. In order to timely deliver the projects, they are being assembled by integrating legacy applications, third party applications, or other components with middleware solutions. But these integration complexities are resulting in project delays. Service Virtualization is an effective means of enhancing the testing process and software delivery, as it supports earlier and frequent testing even when the dependant systems are not available. Service Virtualization more effectively helps to address the various testing challenges in agile development.

Complex systems need to be represented in various development, quality assurance, acceptance and regression test environments. Before we can stand in favor of service virtualization, let us review the various software testing bottlenecks.

- The Test Environment Bottleneck

1. Many systems and applications consist of middleware packages that need software licenses. Hence deploying a testing environment for such a system requires a large number of premium licenses.

2. Setting up and configuring a complex environment requires a number of specialist system experts. These resources are mostly unavailable, as setting and configuring a testing environment is considered as a low priority job.

3. Complex systems also require complicated hardware and operating systems, which makes setting the test environment even more difficult.

- Regression Test Bottleneck: Quality assurance and stability of an application requires a full regression testing while development. However, due to system dependencies, full regression tests can be very weak.

- Common pitfalls: In Agile development, hand written stubs are more in use. This takes time and also these stubs are unable to define the behavior of the real system. This invalidates the entire process of software testing.

- Architectural constraints: Several constraints are encountered while integrating with existing systems. Hence, without the availability of the complete system, the testing process cannot be planned efficiently.

- Continuous Integration Bottleneck: Continuous integration depends on a fully automated system that runs many times a day. Any inconsistency between the components of the existing system can hold back this process.

- Project Planning: The method of project planning can result in a high risk by finding the defects later. Many times due to the reliance on the system with third party components, project planning is most of the times ineffective.

Service Virtualization provides benefits earlier in the lifecycle. In order to virtualize the service, efforts should focus on understanding the interface and the behavior of the service. Testers, developers and architects are able to reuse these assets created during these exploratory work. Virtualized test environments can be used by the multiple users and a large number of assets built around a system and shared by multiple teams. Service Virtualization can resolve the testing bottlenecks in the following ways:

- Virtualize a wide range of resources for a single application

- Virtualize various styles of integration

- Reduce the impact of testing new SOA deployments

- Integration testing of composite services

- Integration testing of automated business processes

- Reduce service dependencies

- Respond more quickly to mobile market needs

Conclusion

For complex heterogenous systems, the challenges of end-to-end testing becomes a bottleneck. The impacts of small changes are exponential when dependant systems are unavailable. Service Virtualization helps in resolving these risks and brings the system integration testing into each sprint. Such solutions help to provide integrated testing and simplifies the test environments for optimized testing and development activities.