Welcome!

Service Virtualization and API Testing

Cynthia Dunlop

Subscribe to Cynthia Dunlop: eMailAlertsEmail Alerts
Get Cynthia Dunlop via: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Related Topics: Cloud Computing, Virtualization Magazine, SOA & WOA Magazine, Cloud Expo on Ulitzer, DevOps Journal

Cloud Computing: Blog Post

Why Leading Companies Love Service Virtualization

For Valentine's Day: A collection of Service Virtualization love stories

Why leading organizations across financial, retail, communications, utilities, travel, insurance, and other industries fell in love with service virtualization...

There are a lot of things to love about service virtualization. Just consider a few of the many exciting findings from voke's recently-released  service virtualization research:

  • 36% of respondents achieved a greater than 41% reduction in production defects
  • 46% achieved greater than 41% reduction in total defects
  • 20% achieved more than 2X the test coverage
  • 26% achieved an increase of 2X or greater of test execution rates
  • 34% achieved a decrease of 50% or greater in test cycle time
  • 40% achieved a decrease of 40% or greater in release cycle time

These stats are quite impressive. But what's a love story without, well, a story?

Here are some great stories about why leading organizations across financial, retail, travel, communications, utilities, insurance, and other industries fell in love with service virtualization...

Why Staples Loves Service Virtualization
Staples is committed to making everything easy for its customers, but ensuring positive customer experiences on their eCommerce site is far from simple. Functional testers must contend with the high number of dependent systems, subsystems, and services that are required to complete almost any eCommerce transaction-but rarely available for dev/test purposes.

The Staples eCommerce functional testing team turned to service virtualization in hopes that it would enable them to more rapidly and more exhaustively test complex transactions across highly-distributed systems.  They found that with service virtualization, they could start testing earlier in each cycle and complete their test plans faster.

This was especially critical on parallel development projects, such as when the Retail, Warehouse, and eCommerce teams were all working on functionality related to online ordering with in-store pickup.  This was a complex project with a very aggressive timeline. Using service virtualization to simulate resources that were still being developed, each team's development and testing could move forward without waiting on the others. With the virtual assets, they could start integration testing much earlier than if they had to wait for all the dependent components to be completed. This helped them get everything running smoothly even before they integrated all the completed components. Ultimately, they not only completed the project on budget, but actually ended up deploying it two weeks early.

In addition to loving how service virtualization promotes faster release cycles, they are also quite fond of how service virtualization gives them greater control over environment stability and application behavior, as well as how it frees them to test on their own schedule.

Read more about service virtualization at Staples

Why Comcast Loves Service Virtualization
Comcast's Performance Testing team often ran into scheduling conflicts around sharing the test infrastructure. Sometimes downstream systems were not available. Other times, test engineers would try to run tests at the same time, which could affect the test results. This led to variability between tests, which made it challenging to isolate particular problems.

When they started with service virtualization 3 years ago, their initial focus was on the biggest pain points in terms of scheduling conflicts within the performance testing teams, unavailable systems, and systems where our testing would impact other development or test groups. Since then, they've been able to virtualize about 98% of the interfaces involved in their tests, and they've seen a 65% annual reduction in the amount of time it takes us to create and maintain test data (factoring in the time they spend creating and updating virtual assets). They also reduced staging environment downtime by 60%.  Their tests are now more predictable, more consistent, and more representative of what would be seen in production. Moreover, they're able to increase the scope of testing in many cases.

In summary, they love service virtualization because it's allowed them to get great utilization from their testing staff, complete more projects on time, and also save money by lowering the overall total cost of performing the testing required for a given release.

Read more about service virtualization at Comcast

Why a Fortune 500 Retailer Loves Service Virtualization
As a leading Fortune 500 retailer advances its omnichannel retail strategy, ensuring a positive user experience on the company's eCommerce site has become increasingly critical. More and more of their customers are now utilizing the ecommerce site at some point during the purchase process- for example, to research products before (or after) visiting a brick and mortar location, to order products for direct delivery or in-store pickup, or even to initiate a product return. Recognizing that all these additional touch points represent opportunities to reinforce-or undermine-their reputation as a market leader, the company is firmly committed to ensuring that all transactions associated with this eCommerce site meet or exceed customer expectations.

They love the fact that service virtualization provides them an efficient and cost-effective way to accelerate the delivery of top-quality functionality. They could rapidly create "virtual assets" for dependencies ranging from mainframes, to SAP, to JDBC, to ESBs, to partner APIs, and countless services- all of which communicate via a variety of message protocols and formats. As a result, all nine regional offices gained anytime, anywhere access to a complete test environment.

Before adopting Service Virtualization, team members would typically wait weeks for access to test data, then try to race through the test plan during highly-limited (and inconvenient) test environment access windows. Now, the team can begin testing as soon as a new service is completed-even if dependent systems are not yet completed or are unavailable for testing-and complete the full range of planned testing. With an unprecedented level of control over the dependencies' behavior, their testing now covers a broader range of "what if" scenarios (e.g., concurrency, fail-over, performance, and negative test scenarios). This extensive, early testing drastically reduces the number of issues that surface when their services are finally integrated into the production system-accelerating the release cycle while reducing business risks.

Ultimately, Service Virtualization has given the team an efficient way to ensure that new ecommerce services are validated extensively and accurately, then are fully optimized before deployment. As a result, the company has been able to reduce costs, accelerate the delivery of innovative new functionality, and achieve their ultimate goal of ensuring a positive, seamless customer experience across the web site, mobile applications, and retail stores. With results like that, what's not to love?

Read more about service virtualization at this leading US retailer

Why Ignis Asset Management Loves Service Virtualization
Ignis recently embarked on a large project aimed at outsourcing the back office as well as implementing the architecture and applications required to support the outsourcing model. A number of projects had to be developed and delivered in parallel, but they didn't have the resources, budget, and management capacity required to create and maintain multiple test environments internally. This limited test environment access impeded their ability to validate each application under test's (AUT) integration with third-party architectures. Moreover, their third-party providers also had limited test environment access, which restricted the time and scope of their joint integration testing.

With an enterprise-grade API Testing solution deployed in concert with leading service virtualization technologies, they were able to reduce the execution and verification time for their transaction regression test plan from 10 days to a half day. This testing is not only automated, but also quite extensive. For example, to test the Ignis system's integration with one business partner's trading system, Ignis's fully automated regression testing now covers 300 test scenarios in a near UAT-level approach-with 12,600 validation checkpoints per test run.

Beyond addressing the original challenges posed by the project, the service virtualization and API testing solution has also enabled automated testing to occur all the way from the component/unit level to system integration. To achieve this impressive level of automation, testers fostered close relationships with the development team. Now, testers' role within the organization is elevated, and there's much more love and collaboration between development and testers.

Read more about service virtualization and API Testing at Ignis

More Service Virtualization Love Stories
Still warming up to the idea of bringing service virtualization into your own organization? Here are even more reasons why so many leading enterprises are now devoted to service virtualization:

  • Financial: Eliminating Third-Party Test Environment Access Fees: A leading brokerage firm needed to increase the scope and frequency of testing without incurring increased fees for accessing their banking partner's test environment. Service Virtualization enabled them to cut their dependency on the partner's test environment by simulating the behavior of partner services (including those communicating over a specialized FIX protocol).
  • Travel: Parallel Development of Highly-Interdependent Components: A global resort group needed to roll out a new heterogeneous, distributed system that involved numerous contractors developing interdependent components in parallel. Service Virtualization allowed the organization to eliminate development deadlocks that stemmed from this extreme interdependency. By virtualizing the expected behavior of "not yet implemented" components across multiple protocols and technologies (JSON, MQ, JMS, REST, SOAP, etc.), the organization enabled each contractor to start developing and testing their assigned components without waiting for dependencies.
  • Insurance: Reducing Infrastructure Costs for Staged Test Environments: An insurance company needed to establish seven distinct test environments for a new application. Each environment had to leverage data from over 20 back-end systems. This was not only complicated, but also costly: licensing the MQ broker that drove communications between the application under test and the back-end systems cost approximately 100,000 Euros per environment. Using Service Virtualization to emulate the interface to the back-end systems, they were able to cut the dependency on those systems and significantly reduce the costs associated with establishing the expected test environments.
  • Utilities: Facilitating Partner Integration & Validation: To establish the infrastructure needed to efficiently transact in a recently-deregulated energy market, this leading energy organization created a new message format and API to streamline communications related to energy delivery and administration. Since the project was on a very strict deadline, partners had to develop and test their integration with this new API at the same time that it was being developed, they simulated the anticipated API behavior using Service Virtualization. They also automated the validation process, lending objectivity and traceability to the partner certification process.
  • Financial: Removing Test Data Management Bottlenecks for Agile: A financial services provider was migrating to Agile to support a continuous delivery model. They soon discovered that completing 2-week scrums on time would be impossible unless they reduced the lengthy wait time for accessing a test environment configured with the appropriate test data. Service Virtualization enabled them to provide immediate access to the necessary test environments; virtual environments with the appropriate data could be set up in hours rather than weeks. Since functional and performance testers can now easily access the same level of sophisticated data much earlier in each cycle, they have been able to expand test coverage and begin continuous regression testing.
  • Non-Profit: Cloud-Based Solution for Continuous Access to a Highly-Restricted Government System: An education portal application developed by a European non-profit organization links students to the higher education institutions where they wish to study, as well as to the government agency that helps them finance their education. When educational institutions want to develop and test transactions involving this portal, they need access to the behavior of the interconnected government agency's system-however, this system is not readily available for testing. Service Virtualization provides these institutions continuous, secure access to the government system behavior that is critical for completing thorough end-to-end tests against the portal application.
  • Telecom: Enabling Faster, Earlier, and More Complete Testing: To accelerate application release cycles, KPN needed to address a critical bottleneck in the testing process. Their end-to-end test scenarios interacted with dependencies controlled by other divisions and external entities, and gaining access to the required dependencies was a slow and frustrating process. Due to these test environment access constraints, testing efforts were regularly delayed and cut short. Service Virtualization enabled them to test faster, earlier, and more completely-enabling fully-automated continuous testing.
  • Government: Service Virtualization in a VMware Environment: At one of NZ's largest government agencies (the Inland Revenue Department), software development procedures had exceeded their architectural limits; projects were taking longer than necessary and costing more due to the serial nature of the software development lifecycle. The impact of this was outages due to issues that should have been resolved during the testing phases earlier in the project. Service Virtualization, working hand-in-hand with VMware, enables testing to commence at a much earlier stage and allows multiple projects to be tested in parallel. Not only does this reduce the time taken to complete projects, it also significantly improves both their quality and reliability

More Stories By Cynthia Dunlop

Cynthia Dunlop, Lead Content Strategist/Writer at Tricentis, writes about software testing and the SDLC—specializing in continuous testing, functional/API testing, DevOps, Agile, and service virtualization. She has written articles for publications including SD Times, Stickyminds, InfoQ, ComputerWorld, IEEE Computer, and Dr. Dobb's Journal. She also co-authored and ghostwritten several books on software development and testing for Wiley and Wiley-IEEE Press. Dunlop holds a BA from UCLA and an MA from Washington State University.