eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Despite the availability of a wealth of development tools and agile methodologies, an alarming 36 percent of the 200 North American organizations surveyed in a recent study found defects in new releases that had gone into production, according to CA Technologies.
In addition, only 4 percent of those surveyed claimed that errors are never found in production releases. This means that many organizations are launching buggy applications to market and having to solve for them later with software updates and patches.
New approaches to testing and modeling software can help address this issue. One example is service virtualization, or the use of a completely virtual environment for software application testing that cuts out constraints or barriers to delivery, said Shridhar Mittal, general manager of service virtualization at CA in a blog post.
Meanwhile, more than half–60 percent–of respondents in the North American ‘Business Benefits of Service Virtualization’ study said customer-facing applications are delayed as a result of endemic constraints within the software development and testing environment, including limited access to infrastructure, databases and undeveloped applications. To compound the situation, applications are often released with reduced functionality, according to 70 percent of those surveyed.
Moreover, the majority of the 200 in-house software development executives and managers from large enterprises—those with revenue of $1 billion or more—surveyed said they are aware of the significant consequences that result from endemic constraints across software development and testing. This includes loss of reputation and customers switching to competitors.
“North American businesses are under pressure to deliver increasingly complex applications, and at a much faster rate than ever before to keep pace with customer demands,” Mittal said in a statement. “Unfortunately, IT budgets are not increasing at the rate of change inherent in today’s highly distributed composite applications. This causes serious constraints to software development, resulting in delays and failures in delivering new software features to market.”
Delays in application development and testing are negatively impacting businesses; 70 percent of the study respondents report reduced functionality and 60 percent report late delivery of new customer-facing applications. In part, this is due to the increased pressure and demand for highly sophisticated applications, with 66 percent of respondents stating that their approach to software development and testing will have to change as a result of massive growth, particularly across mobile.
The pressures highlighted by this independent study point to the need for improved development processes and faster, more effective testing, Mittal said. North American survey respondents also identified the potential benefits of pursuing updated approaches to include increased quality (81 percent), faster time to market (76 percent) and reduced costs (71 percent).
Software Development Failures Plague North American Enterprises: Study
Service Virtualization addresses these challenges by enabling teams to develop and test an application using a virtual service environment that has been configured to imitate a real production environment. This provides the ability to change the behavior and data of these virtual services easily in order to validate different scenarios.
“Service virtualization (or SV) is the automated practice of capturing and simulating any system or service IT teams depend on to deliver software,” Mittal said in his post. “This is not like conventional hardware virtualization that copies some servers to free up hard drives and rack space in your own data center. We are talking about simulating every constraint in the software environment—the very distributed and over-utilized stuff that software teams need to interact throughout development, such as complex customer-response data, core mainframes, integration middleware and performance data that is either costly or unavailable to you.”
“This research follows a European study conducted in July 2012 in which 32 percent of respondents revealed that they were expected to deliver and manage four to seven releases a year, compared to 53 percent in North America,” Ian Parkes, managing director of Coleman Parkes Research, said in a statement. “Even more surprising, 75 percent of respondents across North America and Europe reported they were seeking additional budget to pay for more application development man-hours, when we know that additional labor is not in fact the ideal solution.”
“In short—it’s time for enterprise IT to industrialize simulation and modeling, or suffer more delays and failures,” Mittal said. “While the concept is new for software, it shouldn’t be. Other industries such as aviation and automotive manufacturing have done it for years, using things like wind tunnels, flight simulators and computer models to avoid real-world constraints and do their engineering and testing much earlier, far before real products are assembled. IT needs to apply these same principles of simulation and modeling.”
According to the study, “These survey results suggest that development managers often bring new applications or services from testing environments into production without complete insight into how their integrated applications might fail. For engineers, understanding failure modes is a critical part of the job, yet according to this study, 69 percent did not have this insight on a consistent basis. This is an alarming prospect for any board giving the green light for new software projects, especially those that impact the customer. It is also concerning that only nine percent have comprehensive insight into how complex, integrated applications could break in production.”