Getting the Mainframe Up to DevOps Speed
January 23, 2017

Chris O'Malley
Compuware

Until recently, many IT leaders still believed they could allow their mainframe environments to languish in two-code-drops-a-year waterfall mode, while they embraced DevOps and Agile across their distributed and cloud environments.

This so-called "Bimodal lT" strategy has proven to be dangerously flawed. The fact is, if your business has a mainframe, that's probably where your most important applications and data live. As such, there's no way your business can remain competitive unless you can quickly adapt your use of those applications and data to keep pace with your rapidly and relentlessly evolving market demands.

That's especially true given the fact that your customer-facing mobile and web systems of engagement almost universally leverage your back-end mainframe systems of record.

So how do you actually get your mainframe environment up to speed? Given the fact that your existing mainframe dev/test processes and tools are pretty entrenched, how can you integrate the platform into a truly nimble and unified cross-platform enterprise DevOps environment?

Different organizations will take different approaches to this challenge. But here are three principles to bear in mind as you go about the difficult but ultimately extremely rewarding work of bringing your mainframe into the DevOps fold:

1. Transform the developer workspace

Most mainframe dev, test and ops work is still performed in "green screen" TSO/ISPF environments that require specialized knowledge, constrain productivity, and are extremely off-putting to the kind of skilled, ambitious programmers who are the lifeblood of Agile and DevOps transformation. It is therefore essential to migrate to more modern, graphical tools within a preferred DevOps toolchain that empower staff at all experience levels to perform mainframe tasks in much the same manner as they do other non-mainframe work.

Also, mainframe applications are typically large, complex and poorly documented. These attributes are a major impediment to mainframe transformation — and they tend to make enterprise IT highly dependent on the personal/tribal knowledge of senior mainframe staff.

To overcome the skills-and-knowledge gap, it's not enough to just make mainframe workspaces more graphical. You also need tools that enable new participants in mainframe DevOps to quickly and easily "read" existing application logic, program interdependencies, and data structures.

Recent innovations in mainframe workspace technology can also give developers on-the-fly feedback on any new bugs and quality issues they inject into their code. By investing in these tools, IT can empower even mainframe-inexperienced developers to quickly produce quality work that fits within the daily requirements of an Agile process. In addition, the latest mainframe development dashboard solutions enable managers to track defects, program complexity and technical debt so they can better pinpoint issues requiring additional coaching or training.

2. Remodel mainframe processes

Once you've built a better working environment for the mainframe, you can start to aggressively shift your process from a traditional waterfall model with large sets of requirements and long project timelines to a more incremental model that allows teams to quickly collaborate on so-called user "stories" and "epics." By estimating the size of these stories and assigning them their appropriate priority, your teams can start engaging in scrums that allow them to quickly iterate towards their goals.

The move from large-scale waterfall projects to Agile scrumming represents a significant change in work culture for most mainframe teams. Training in Agile process and work culture is therefore a must. You may also want to build your initial Agile mainframe team by choosing select mainframe developers and pairing them with Agile-experienced developers from other platforms to work collaboratively on user stories and epics.

You'll obviously also need the right enabling technologies for this shift. Key requirements include Agile project management software that supports Agile methodology — as well as Agile-enabled Source Code Management (SCM). The latter is especially pivotal since traditional mainframe SCM environments are inherently designed for waterfall development, and are thus incapable of providing essential Agile capabilities — such as parallel development work on user stories.

When engaged in this re-tooling, it is generally wiser to leverage best-in-class tools rather than fall into a monolithic approach that requires all SDLC activities to be performed within a single vendor's solution set. That's because best-in-class tools allow you to avoid vendor lock-in while taking advantage of the latest innovations in Agile management.

3. Integrate mainframe workflows into the cross-platform enterprise DevOps toolchain

The target state of mainframe transformation is ultimately a de-siloed enterprise DevOps environment where the mainframe is "just another platform" — albeit an especially scalable, reliable, high-performing, cost-efficient and secure one — that can be quickly and appropriately modified as needed to meet the needs of the business by whoever is available to do so.

This requires integration between mainframe and distributed tools (typically via REST APIs) so that DevOps teams have a single point-of-control for all changes across z/OS, Windows, Unix and other platforms. An effective cross-platform toolchain will also provide cross-platform impact analysis — so your developers can see how the code they're working on in one tier of an application (e.g. a mobile app server) may potentially affect another application tier (e.g. a DB2 database).

The de-siloing of your mainframe can also lead to unified IT service management (ITSM) for both mainframe and non-mainframe applications. This unified ITSM model is for companies with large numbers of multi-tier applications that are critical to their financial performance.

Of course, it takes budget, hard work and strong leadership to turn these principles into in-the-trenches realities. It is important for mainframe users to align themselves with others that don't just pay lip service to bringing the mainframe into the DevOps fold, and are committed to actually doing what it takes to get it done, and have experience doing it within their own organizations. Getting the mainframe up to DevOps speed is possible — and is being done by enterprises that recognize the need to complement their existing advantages of scale with new advantages of speed. And the alternative of "bimodal," "two-speed," or "multi-speed" IT is simply untenable. 

Every company with a mainframe therefore needs to get with the mainframe DevOps program. Now.

Chris O'Malley is CEO of Compuware.

The Latest

February 15, 2018

Organizations are focusing investments on security and analytics, while actively researching emerging technologies such as machine learning and artificial intelligence, according to the 2018 CIO Tech Poll: Tech Priorities survey ...

February 14, 2018

With so many other initiatives to manage, DevOps isn't a top priority for all companies now. Many organizations believe it's wiser to wait. But in fact, the business case to integrate now is very strong — in fact, it's "do or die" territory. This infographic outlines some key reasons why organizations should integrate their software delivery value stream immediately ...

February 12, 2018

Accelerating multi-cloud deployments are enabling organizations to select the cloud platform that best meets the requirements of a specific application, according to the 2018 State of Application Delivery report from F5 Networks. However, this also increases the challenges many companies face in managing operations and security across multiple clouds as they transform their application portfolio to compete in the digital economy ...

February 08, 2018

The slowness of enterprise IT departments to embrace automated, cloud-native solutions for the cloud infrastructure challenges they face has resulted in IT infrastructure that is often ungoverned and insecure. And this is despite the fact that the cloud can be more secure as traditional data centers ...

February 06, 2018

Bank IT teams must embrace that their DevOps capabilities will determine their agile capability. Agile breaks down the barrier between the business and IT, and operations must be treated as a critical element of an agile program. In modern software delivery, the business, development and operations must execute as a unified team. To achieve this, banks are increasingly turning to Continuous Integration (CI) practices as part of the solution ...

February 05, 2018

IT professionals show a heightened concern for cybersecurity risk related to API use, according to a new survey conducted by Imperva. Specifically, 63 percent of respondents are most worried about DDoS threats, bot attacks, and authentication enforcement for APIs ...

February 01, 2018

DevOps are pretty clear for application development, those same applications often have a database back-end. If DevOps is increasing the frequency and reliability of new features for applications, a slower pace of database development can slow down and hinder those same releases ...

January 30, 2018

Without a doubt, DevOps is becoming the go-to strategy for organizations of all industries and sizes looking to master digital transformation and provide the fastest value to customers through software delivery. It is becoming clear that organizations adopting DevOps need a true leader (or engineer) to keep the transformation on track. The following are some tips when hiring for DevOps ...

January 29, 2018

When was the last time your company experienced a significant database error? If it happened in the last 3 months, you’re in good company. In a recent study, 60% of respondents reported a crash or significant database error occurring in the last 6 months. Roughly one in ten respondents reported a serious database problem in the past week ...

January 25, 2018

You've already recognized that business transformation requires digital transformation. Your organization is staffed with the best and brightest developers ready to implement the innovative, business-differentiating technologies you need to attract, engage, and retain customers. And you've invested in scaling Agile, driving DevOps adoption, automating the Continuous Delivery pipeline, and all the other components involved in moving from ideation to delivery as rapidly as possible. So what could possibly go wrong? Testing ...

Share this