13/05/2020 | By Vincent Amet and Laurent Coutellec, Practice Managers, IS Transformation, Hardis Group

At one point, we thought we’d seen the end of AS/400-IBMi systems, sacrificed on the altar of modernization. But that view underestimated just how robust, reliable and secure these systems can be—and how they can create business value. Here, we examine why legacy systems remain central to the digital transformation of business, and what role they will play.

AS/400-IBMi: an unexpected reversal of fortunes

Not so long ago, the general consensus was that AS/400 systems were on their way out because they were outdated, complex, costly (especially in terms of licenses), and unfit for the Internet age. 

Anticipating the coming change, some companies set about replacing their legacy systems. But the scale and complexity of the endeavor often proved daunting, not least because they had dozens, or in some cases hundreds of thousands of lines of undocumented code, as well as considerable technical debt. In the end, many opted to replace their applications on a like-for-like basis instead—often with better outcomes.

Given the eye-watering costs associated with rebuilding almost entire systems from the ground up, some companies have tried automating the process of rewriting applications, albeit with limited success. Others have embarked on time-consuming ERP integration projects.

Yet for all these initiatives, as of 2020, IBM figures indicate that there are still 100,000 companies worldwide using legacy AS/400 systems, including between 3,000 and 4,000 in France alone. The environment has by no means gone away. And it still forms the core of many business applications on account of its reliability, robustness and security. What’s more, IBM has continued to upgrade its platform with additional capabilities, including compatibility with emerging technologies. AS/400 has become IBMi.

As a result, more and more companies are opting to retain this legacy system—so much so that IBM has published a product roadmap running to 2032.

A business-application back-office and the cornerstone of digital transformation

Because legacy applications perform business-critical functions, they have naturally evolved with the passage of time. Using IBMi-specific computer-aided software engineering (CASE) tools, businesses have, for instance, been able to retain their legacy systems while adding new features and limiting the risk of technical debt. But as support has been withdrawn for some of these CASE tools, and others have been retired altogether, businesses are finding it hard—if not impossible—to maintain their legacy applications. In cases like these, businesses have no option but to convert their software to IBMi native language. Interestingly, some vendors have released automatic code conversion tools, in free-form RPG format, allowing businesses to give their legacy applications a second lease of life.

Meanwhile, businesses have developed fully fledged strategies in an effort to strike the right balance between offering new services (by extracting value from business-application back-office data) and keeping a tight rein on development costs. Three key rules underpin the modernization of AS/400-IBMi legacy information systems.

The first rule is to develop modern application interfaces (user interfaces, or UIs) in order to improve the user experience (UX), while recognizing that, in some cases, the old-style green and black screen is still perfectly adequate and makes it easier to input information quickly.

The second rule is to shift to modular applications (adopting a Service-Oriented Architecture or an Event-Driven architecture): this allows businesses to design new, flexible digital services, such as UIs developed using the latest languages and frameworks. In this case, the legacy system takes care of core business functions. Such an approach requires careful development planning of the target system, including the use of APIs and middleware modules more generally.

The third and final rule is to adopt continuous integration principles—a standard DevOps software development practice—so that, once integrated into a modern information system, the AS/400-IBMi legacy system has its own maintenance and development cycles. 

Auditing legacy systems to make informed decisions

Although many companies are seriously considering maintaining legacy systems, it remains the case that information may have been lost or forgotten over time. For this reason, businesses need to carry out an exhaustive application mapping exercise, or update their map if they already have one, focusing first and foremost on business-critical systems. They must also audit these systems constantly to ensure they remain viable. 

Business teams will need to be involved in this process, and the company will need to seek input from people responsible for maintaining and upgrading IBMi applications in the past. Aside from holding these conversations, businesses should employ tools that automatically identify database architectures and conceptual data models (CDMs), platform types and compatibilities between them, hardware configurations and hosting arrangements, and more.

Only once they have completed this exhaustive audit exercise will businesses be able to make informed decisions around data governance, and internal and/or external team organization, as well as whether to opt for public, private or hybrid cloud models for greater scalability. 

During the audit, the business can also consider whether it might make sense to switch some applications over to a software-based solution. There is a vast array of software out there on the market covering an increasing range of use cases, especially for support functions. 

Despite their age, legacy systems appear to be achieving something remarkable: staying relevant in a digital world that many thought would leave them behind. But as many experienced RPG programmers have retired, and many more are planning to do so soon, businesses are likely to face yet another challenge: attracting the talent they need to keep these systems running.