carloscastilla - Fotolia

Digital transformation needs mainframe DevOps

As organisations become more digital, CIOs must consider how best to tap into data held in mainframes

Mainframe systems of record are the beating heart of most large businesses. The mainframe is essential to 92 of the top 100 banks worldwide, 23 of the top 25 US retailers, all of the world’s 10 largest insurers, and 23 of the world’s 25 largest airlines.

Successful digital businesses unleash the data and business processes encoded in their mainframe-based applications.

Ad hoc integrations between systems of engagement and systems of record get them started, but they soon find their ability to define innovative products and services is limited by an inability to evolve and improve their mainframe applications.

Mobile and cloud applications garner the greatest attention, but nearly every aspect of what they do depends on data and processes locked in mainframe apps. A seemingly simple mobile insurance app is actually the gateway to a complex set of applications that must work seamlessly with the app and with each other.

Many companies speak of their mainframe-based applications as assets, but they are more often liabilities that impede their ability to innovate. Those old applications embody calcified processes that prevent companies from responding to market opportunities with new products and services.

IT leaders have legitimate fears about modifying applications they know little about. Combinations of code analysis tools and services help them better understand the structure of their applications, letting them make informed changes with greater confidence.

The various approaches to this have different advantages and disadvantages: service-based approaches help IT organisations quickly eliminate technical debt. Offerings such as those from Infosys and The Software Revolution use analytic tools plus expertise to help IT teams make sense of the mess.

Mainframe applications are complex. While tools provide insight, human judgement is often needed to make the right call about where and how to restructure and modernise.

These services make the most sense for organisations undertaking significant restructuring, re-platforming, or rewriting of specific mainframe applications.

Low risk insight into application code

Tools-only methods favour a gradual self-serve approach. For organisations keen to do their own analysis, tools like Cast, Compuware’s Topaz and IBM’s Rational Asset Analyzer provide insights into application code that increase understanding and reduce the risk of change. These tools help them understand the structure of the code and identify opportunities for improvement.

Static code analysis helps developers improve code quality. Integrating static analysis, using tools like RogueWave’s Klocwork and SonarQube, into pre-commit processes helps to correct modularity and code quality problems before they affect the main code base.

Developers tend to reject static analysis when it is used to judge or evaluate them or when the rules are poorly defined and produce too many warnings on non-critical errors.

Reducing thechance of code change

Dynamic analysis and analytics help pinpoint operational issues. Instrumenting and monitoring mainframe applications, as well as moving production monitoring tools and practices into testing environments, also help developers reduce the risk of changing code. Identifying and resolving issues earlier reduces the chance that changes will lead to problems in production; when they do occur, the issues can be found and resolved faster. Tools such as IBM Application Delivery Intelligence and Compuware Strobe are being integrated with widely used DevOps tools from companies like AppDynamics, Atlassian, Jenkins, SonarSource and Splunk to provide greater insight across the life cycle.

Antiquated coding tools such as ISPF present significant barriers to younger developers working on mainframe application code. Familiar tools, such as IDEs with their integrated debuggers, syntax-directed editors and familiar interfaces, make picking up Cobol or PL/1 relatively easy. Tools such as Compuware’s Topaz Workbench, IBM Rational Developer for zSystems, and Micro Focus Enterprise Developer eliminate the need to learn old platform-specific tools and provide rich developer experiences that bring modern practices to bear on old code.

Realising the benefts

Mainframe applications must not prevent organisations from reaching digital transformation goals. The same modern DevOps practices that help IT organisations deliver mobile and cloud apps faster will help them deliver mainframe applications faster, too.

To start realising these benefits, IT leaders should only modernise mainframe applications that are essential to digital experience. Most organisations have vast portfolios of mainframe applications. They can’t all be modernised at once, nor do they need to be. Focus on the applications that provide information critical to customer experiences.

These applications provide essential data to customer journeys, define and manage the products customers buy, or manage the processes that guide the customer journey.

Read more about DevOps

Forrester recommends funding mainframe application improvements through digital initiatives. Organisations struggle to fund modernisation for modernisation’s sake, but they usually have plenty of money for customer-facing initiatives that create new revenue opportunities. Clearly delineate how improving mainframe application delivery is essential to the success of digital initiatives, and build funding into the digital initiative budget to make the necessary improvements.

Forcing developers to use antiquated tools and processes that only apply to mainframe development turns new developers off. Giving them modern and familiar tools makes mainframe development just another technology to master and reduces the amount they need to learn, so they can be productive much faster.

While languages are different, developers are generally multilingual and adopt new languages all the time. The fundamental processes are exactly the same, as are the benefits of automation and standardisation. Improve mainframe application delivery speed by using automation to speed, simplify and standardise the process.

The mainframe is not going away, but the way it is used will change. Containers and microservices are coming to every platform, including the mainframe. Breaking large monolithic applications into smaller services will help the transition to a containerised future that promises faster application delivery, greater scalability and better manageability.

This is an extract of the June 2016 Forrester report: “Digital transformation needs mainframe DevOps,” by principal analysts Kurt Bittner and Robert Stroud.

Read more on Software development tools