DEVOPSdigest

Industry experts offer thoughtful, insightful, and often controversial predictions on how DevOps and related technologies will evolve and impact business in 2020. Part 6 covers testing.

Start with 2020 DevOps Predictions - Part 1

Start with 2020 DevOps Predictions - Part 2

Start with 2020 DevOps Predictions - Part 3

Start with 2020 DevOps Predictions - Part 4

Start with 2020 DevOps Predictions - Part 5

AI-BASED TESTING

AI-based testing will become more popular as companies continue to focus on the nuances involved in digital transformation. It will thus become critical for Leaders to continuously develop strategies for intelligent testing that will improve customer experience across the board.
Raghuram Krovvidy
EVP, Global Delivery, Cigniti Technologies

It won't be a total takeover in 2020, but we're moving irreversibly in the direction of AI-driven test authoring, in which organizations that may not have the expertise or bandwidth to author sophisticated test suites begin to lean more heavily on AI solutions that automate the initial stages of work. It won't all happen at once, as most of the enabling technologies are still in their nascent stages, but we'll eventually look back on 2020 as the year AI started to take over test authoring.
Joanna Schloss
Evangelist, Sauce Labs

Maybe this will be the year when we stop hearing people argue about the benefit of development and test automation. Modern tools that augment developers' capabilities with AI and machine-learning improve the process, take some of the drudge out of the system, and overall make developers more effective. In particular, testing is an area that is tedious, repeatable, and not well-liked by most developers. New tools that are leveraging AI and ML are helping to create tests that are more maintainable by creating smarter tests than was possible in the recent past as well as helping understand test results more quickly. This frees up precious development resources to apply their human intelligence to solving the difficult problems that automation cannot do.
Arthur Hicken
Evangelist, Parasoft

TWO PATHS TO AI-BASED TESTING

The AI/ML conversation will mirror the early DevOps conversation. As AI/ML becomes more prominent throughout 2020, the discussion of how best those technologies should be applied will mirror the early DevOps conversations with factions emerging. Ultimately, there will be two trains of thinking with some companies implementing full-on autonomous testing and others taking a more human-led approach to AI that helps accelerate productivity by merging machine and mind. As this conversation progresses, we'll see a push and pull between these two paths of implementation as companies think about how these technologies either replace or augment the teams they have today.
Eric Sargent
VP of Sales, Functionize

AUTONOMOUS TESTING NOT REALITY YET

We know a few solutions claim to do autonomous testing; generating tests automatically based on web traffic or common user flows, but are they ready for primetime? Understanding how traffic flows through a website is important and it can help inform development and operations where to focus their energy, but the resulting tests are either too basic, too unstable, or require major customization to be useful. In 2020, AI will absolutely have a significant role to play in testing. However, the ability to monitor traffic and convert that to a meaningful, stable test that eliminates the need for testers is still a few years away.
Shawn Jaques
Director of Product Marketing, Testim

INTEGRATION OF MANUAL AND AUTONOMOUS TESTING

Even as the use cases and testing types for automation expand, there is still a large need for manual testing. They are used in tandem in order to increase release velocity for Agile teams, but, for the most part, test automation and manual testing remain separate from one another. That changes in 2020. I think we'll see companies begin to integrate manual and automated testing into a single view so that bugs, statuses and reports from both can be seen in one unified infrastructure, which will assist in increased release confidence and throughput.
Jonathan Zaleski
Head of Applause Labs

TESTING EXPANDS RIGHT

The concept of "expanding right" grows as performance and quality testing is gaining prominence at all stages of software development. Testing now begins with developers and continues through pre-production into production environments, driving an interest in blurring the lines and testing apps "as if in" production. New approaches to continuous testing allow actionable context and precise diagnosis of software failures that existing tools do not adequately provide, and in 2020, I believe these new approaches will continue to emerge.
Robert Ross
CTO, Curtail

TESTING SHIFTS LEFT

2019 brought about a slew of headline-making software outages from some of the biggest companies in the world, ranging from multiple major banks and airlines to technology giants and startups alike. The exponential costs of downtime are difficult to ignore, as is the growing pressure to outpace competitor innovations. In response to this, shift-left testing has emerged as a way to place greater emphasis on quality earlier in the SDLC — i.e. before production where the implications can be catastrophic — without reducing thee speed of delivery. In 2020, we will see more organizations adopting these principles. We will start to see best practices and better defined processes emerge as shift-left early-adopters not only fine tune their approach, but offer up quantifiable case studies on how shift-left has positively impacted their application reliability.
Tal Weiss
CTO and Co-Founder, OverOps

As software teams look to ship faster, learn and iterate, they are held back by end-to-end (E2E) testing that is either too slow or too flaky. Shift-Left is not a new concept, but for QA, it has primarily been limited to unit and integration testing, and only 32 percent of teams are automating E2E tests. However, due to the increased use of open-source and 3rd party components to build software, I believe that E2E regression testing is more important now than ever. In 2020, expect to see E2E test automation increase, especially early in release cycles. This is aided by solutions that leverage artificial intelligence (AI) or machine learning (ML) to create stable tests faster. Coded tests that leverage AI/ML will gain traction as they can be stored in version control next to the application code, resulting in more immediate feedback on every commit.
Ronit Belsen
COO, Testim

We will see organizations continuing to automate performance testing processes and "shifting left" to identify problems in application code earlier. As the mainframe workloads increase and continue to be highly relevant, we expect this trend to become increasingly evident with COBOL code, helping improve overall application response times while also reducing the demand that apps make for system resources.
John Dannenberg
Product Manager, Compuware

SERVICE VIRTUALIZATION

Too many organizations are still ignoring or underutilizing performance testing. Performance is an important issue of quality and end-user experience. To meet the modern agile and DevSecOps team's performance testing has to shift left along with the other activities like security and functional testing. I believe we'll see the resurgence of service virtualization as a key enabling technology that will allow organizations to start performance testing before all components are ready, as well as allow them to try out scenarios that would be difficult, expensive, or even impossible to perform against a real production system. Service virtualization will improve performance, quality, and even security as part of the shift-left efforts.
Arthur Hicken
Evangelist, Parasoft

NON-FUNCTIONAL TESTING

The evolution towards "Shift Left" test automation has won over the past decade, and most organizations have embedded functional test automation within their CI and DevOps pipelines. In 2020, organizations will integrate more and more non-functional testing into their pipelines, focusing on performance testing that assess the end-user experience, software response time and availability, as well as security of their digital apps. On one hand, markets will become more and more fragmented, and advanced technologies such as 5G are becoming standard, while on the other there is a growing pressure on software development teams to ensure high performance and availability of apps. With that, having the ability to assess performance quality at any given stage of the development life cycle will become a critical requirement.
Eran Kinsbruner
Chief Evangelist and Author, Perfecto at Perforce Software

ATOMIC TESTING

Ideal for the early stages of development, an atomic test is one that's scripted to assess just one single piece of application functionality, and nothing more. Atomic tests are short, fast, reliable, and incredibly effective. They're also vastly underutilized, a trade secret of sorts, championed by only the most experienced and knowledgeable testers. But that's about to change, as 2020 is the year everyone gets in on the secret. The more developers take control of testing, and the more early-stage testing becomes a staple of the software development cycle, the more critical and widespread atomic testing will become.
Joanna Schloss
Evangelist, Sauce Labs

ACCESSIBILITY TESTING

2020 will be the year accessibility becomes a first-class citizen. There are plenty of reasons to focus on digital accessibility — from a moral standpoint to a financial one, not to mention that it is extremely hard to build in after the fact. The issue has been boiling just beneath the surface for years now, and 2020 will be the year it is forced to the top. I believe we'll see accessibility testing shift left in the development process, making it harder for engineers to ignore, and forcing teams to make fixes before they become major issues (i.e. lawsuits, technical problems and disenfranchised users) down the line.
Jonathan Zaleski
Head of Applause Labs

Recent decisions from the SCOTUS underscore the fact that digital accessibility is a business requirement. Digital accessibility lawsuits will continue to skyrocket in 2020. Organizations that embed accessibility testing into their DevOps practices not only put themselves in a defensible legal position, but do so in the most effective an efficient way available. Strong DevOps shops will continue to differentiate themselves in this way.
Dylan Barrrell
CTO, Deque Systems

DEVELOPERS TAKE OWNERSHIP OF TESTING

One of the areas where developers' growing autonomy will most clearly be seen is testing. We've already seen organizations take steps in this direction in recent years, but in 2020, we'll see a true sea change in which developers begin to fully take ownership of testing. Developers no longer see themselves as simply responsible for writing code. Instead, they rightly see themselves as builders and gatekeepers of the entire application experience. In the digital era, developers are ultimately the ones responsible for rapidly delivering a high-performance, high-quality digital experience to customers. And that means they need to assert greater ownership over testing. In 2020, we'll see the shift-left movement accelerate, with developers increasingly seeking the autonomy to run early-stage tests themselves, and increasingly demanding that organizations put a test strategy into place before they begin the work of coding.
Joanna Schloss
Evangelist, Sauce Labs

FOCUS ON TESTING EXPERTISE

Despite the collective investment in DevOps and agility, software release velocity has actually started to stall out in recent years, with testing continually cited as the most common source of development delays. The problem isn't a lack of technology; there are plenty of automated testing solutions that address tooling. The problem is lack of expertise. Achieving automation at scale is hard. It takes a level of experience and know-how that tooling simply can't abstract. Needing to kick-start stalled release velocity, organizations will increasingly look to invest in testing expertise in 2020, either by hiring resources directly or calling on their vendor partners to deliver expert resources as part of their solution offerings.
Joanna Schloss
Evangelist, Sauce Labs

More 2020 predictions to come, covering DevSecOps, containers and the cloud.