Feb 09, 2016 – HfS Research, a research body that provides analyst coverage of the best practices and innovations for enterprise services professionals, sat down with the executives of Undo to show how the company is advancing the automation of services within the testing industry. See the original article or read on to find out more.
By Tom Reuner
Testing services are rarely part of broader strategic IT discussions. Instead, they are generally a secondary item on major transformational projects. The reasons for that are manifold: preconceived views on the benefits, a lack of marketing investment, immaturity of organizational models, a highly fragmented supply-side, and inadequate marketing top the list. For organizations that aspire to leadership positions in the As-a-Service Economy, these historic attitudes toward testing need to change. In an environment where services are being consumed in a plug-and-play modus, testing has to be a primary consideration.
We recently sat down with executives of Undo, a UK-based testing tool provider, to discuss how its approach is advancing the immediacy of testing services and helping to respond to the new realities. HfS came across Undo because it was among 15 startups selected by Accenture’s FinTech Innovation Lab as part of its mentoring program. The company has developed a unique record, rewind and replay technology that developers can use to see exactly what their program did at every step in its execution. Developers can then use reversible debugging and offline analysis to fix problems fast. Its products help those involved in software testing to respond quickly to failures in test environments, to improve the quality of their in-production software and to increase developer debugging productivity by—it claims—50%. Undo does this is through advancing automation within the testing services.
The Undo engine provides deterministic recording of software behavior, allowing execution to be replayed, at a later time or on another host. Using an optimized JIT compiler to instrument unmodified code, the Undo Software technology captures the information required to replay the behavior of programs running on Linux and Android operating systems. Furthermore, it supports all compiled languages, including C/C++ and Fortran. This information can be used to debug an active process, or can be saved to a file for later analysis. The Undo Engine is a reverse debugger. It is used in two key products: Undo DB, the main debugger, and LiveRecorder, which removes the need to write test cases by debugging an exact copy of faulty code. What might sound very technical boils down to the ability to record, rewind and replay the development of code. As Undo Software executives put it, one way to think about the value proposition of its products is that it’s like having CCTV capabilities for code development, with the simplicity of having a play and rewind button. This also calls out the key differentiation: While other tools offer the ability to play and continue, UndoDB offers a record and rewind button.
The discussions with Undo reminded HfS of the early robotic process automation (RPA) vendor briefings with their focus on recording, scheduling and executing process steps in a drag and drop way. Similar to the nascent development phase of RPA then, Undo is in a very early stage. Another parallel to RPA is that the ability to record is the basis for deep analysis of process steps and code, respectively. An example for this is that both tools extract information that can be used, for instance, for compliance issues. Deep analysis of code will also advance product engineering in the transition to new architectures, such as from x86 to ARM. This will be further enhanced by Undo's plans to support Cobol and Java within the next 12 months.
HfS expects that test automation will come to the fore and drive more holistic approaches to Intelligent Automation. Aligned with the notion of the Intelligent Automation Continuum, test automation will play out in many different facets. While Undo Software cannot overcome the systemic challenges of the testing community by itself, it provides an important building block to advance testing services toward the emergence of the As-a-Service Economy.
The interplay of cognitive computing, artificial intelligence and test automation is a central theme on the research agenda of HfS’ IT practice and we will include a comprehensive look at the automation of testing in our 2016 Blueprint on Application Testing later this year. Should you be interested in discussing other innovative approaches, do get in touch: firstname.lastname@example.org