The key to mainframe DevOps success is in quickly identifying and removing major bottlenecks in the application delivery lifecycle. Major challenges include collaboration between mainframe and distributed teams, lack of visibility into the impact of software changes, and limited resource flexibility with scaling out necessary testing initiatives. Now let's take a closer look at some of these key challenges and how IT departments can address them ...
DevOps experts — analysts and consultants, users and the top vendors — offer thoughtful, insightful, often controversial and sometimes contradictory predictions on how DevOps and related technologies will evolve and impact business in 2017. Part 2 covers development and testing.
Start with 2017 DevOps Predictions - Part 1
10. SHIFTING LEFT
I expect standardization of environments to become more important, as the emphasis on "shift left" continues and organizations realize the value of exposing developers and testers to holistic production environments up front via concepts like sandboxes, which become more apparent in speeding up the delivery of software releases at scale.
11. SOFTWARE ACCOUNTABILITY
2017 will be the year that "software accountability" comes firmly onto the agenda. As software continues to saturate the world, more aspects of our lives will depend upon it, both inside and outside of the office. Knowing what software has really done, especially why it has failed or misbehaved is going to be key. Currently, identifying software failures consists mostly of using extensive brute force to log sensor data and decisions made, but the sheer volume of data means we are going to need to automate the recording of software's behavior. That recording will be most effective across all of DevOps, from development to the testing process and all the way to deployment and full production.
CEO and Co-Founder, Undo
12. THINKING IN TERMS OF "PRODUCTS"
As DevOps teams assume greater responsibility for the full lifecycle (to "build and run") of their applications, the trend of treating their systems as "products" will continue to gain traction. This will result in disruptive decisions about planning and budgeting, resource allocation, and alignment to business objectives. Simply stated, if a product team can deliver small releases every week or faster, how IT plans and manages teams will have to change.
Head of the Software Digital Research Team, Hewlett Packard Enterprise
13. IMPROVED DISPUTE RESOLUTION BETWEEN TEAMS
Performance (i.e. speed, reliability) issues have traditionally been a lightning rod for conflict within even the most senior DevOps teams – neither the "dev" or the "ops" side of the house is eager to take the blame, leading to stalemates, conflict, and wasted time and effort. We expect DevOps teams will adopt more clear-cut guidelines for resolving performance disputes, as well as establishing "line in the sand" types of consequences. For example, "If this application does not deliver five nines availability, no further software will be put into production." This is tough medicine, but it incentivizes all teams to work together to enhance performance, expeditiously and collaboratively.
CEO and Founder, Catchpoint
14. AGILE CONTINUES
In my opinion the biggest trend in software development for 2017 will be the continued shift to agile based software development methodologies. Our current social media driven society is pushing organization to release new feature filled web and mobile applications at a lighting pace. Traditional Waterfall software development methodologies are too slow to support the "always on" requirements of today's consumers and businesses. If an organization does not provide the latest and greatest web or mobile application people will simply look elsewhere.
Global Director Application Security Strateg, Checkmarx
15. APP DEVELOPMENT AT LIGHT SPEED
Speed of Application Development: In order to stay competitive organizations have speed up their application development to light speed by moving away from the traditional three layers of testing to an automated model. Unfortunately, some organizations hesitate to automate and continue this methodical approach to application development resulting in competitors eating up any market share available. Some of today's most advanced applications are the ones that are integrated and automated within the test automation phase. Prediction: We predict that more and more automation will require new levels of testing to speed up the development process. Testing and analytics tools today can provide a holistic view of application development to where you can now test new features that weren't available months ago.
Founder and CEO, Apica
Read Sven Hammar's blog: 6 Tips to Avoid "Integration Hell"
In response to higher business demands for agilty and feature velocity, cutting-edge software development shops will start matching Amazon's release benchmarks of several production-ready changes per minute. Run-of-the-mill shops also become more agile, but their production release rates will be measured in days, not minutes.
Co-Founder and VP of Products, Aporeto
In 2017, DevOps teams must focus on bringing value to customers faster. This kind of innovation at hyperscale requires a "continuous delivery production mindset" and an architecture that will avoid war rooms. Getting this right means developers must think and plan for monitoring before they even write the first line of code so they get the right feedback from production systems – to drive for desired business outcome in an agile way. The goal is a "one hour code to production" framework to bring business value to customers in a rapid fashion, through frequent releases or continuous delivery.
SVP, CTO, Dynatrace
16. HOT BUZZWORDS: PIPELINE AND RELEASE ORCHESTRATION
Pipeline and release orchestration will be the two hottest buzzwords for the non-Dev folks in the application delivery process, as these approaches provide the consistency in the delivery process that business demands, along with the flexibility that IT teams need to deliver solutions faster.
VP of Customer Success, XebiaLabs
17. LOW-CODE AND NO CODE PLATFORMS
DevOps will start to leverage low-code and no-code platforms to shorten the development cycle from weeks to days. Forrester defines low-code platforms as those which "enable rapid delivery of business applications with a minimum of hand-coding and minimal upfront investment in setup, training, and deployment". This does not eliminate the DevOps cycle, but greatly accelerates it and allows IT organizations to keep applications fully up to date with business requirements. IT will finally get the respect it deserves.
18. CONTINUOUS TESTING
Continuous Testing will be a big part of DevOps 3.0. Companies adopting DevOps will need to improve their ability to test and perform test automation. As part of that, test acceleration will be a key area of investment – as you start getting more and more automated test suites. The time it takes to run those test suites can be very long, so you'll see ways of implementing some test acceleration capabilities that can radically reduce the amount of time for some of those tests.
GM, Acceleration Solutions, Electric Cloud
19. NETWORKED APPLICATION TESTING
DevOps teams will recognize that if they are to deliver meaningful application development and testing they are going to have to "go the extra mile" and broaden the scope of their testing processes to embrace Networked Application Testing. It will no longer be regarded as satisfactory for the DevOps team to simply stop the testing process once they have completed functional testing. Networked Application Testing will provide the team with a complete understanding of how the network, over which most applications are now delivered, will impact the user experience. This is essential, because in our highly connected world where users can share their experiences (good or bad) almost immediately, businesses will need to "know" that their apps will work effectively over the network, rather than "hope" that they do. The potential consequences of failure, such as reputational damage and loss of revenue, are too high. To mitigate such failures, DevOps teams will come to regard Networked Application Testing as an essential part of the DevOps process and not a "nice to have".
Operations Director, iTrinegy
Read 2017 DevOps Predictions - Part 3, all about the cloud.