The 5 Myths Of Mainframe Modernization

Photo Credit: Shutterstock/ Server room in datacenter

As cloud-based operations, mobility, and big data have become household words in the business world, older enterprises have had to play catch up with newer businesses not saddled with legacy mainframes. The challenge is that modernizing legacy applications, either by extending compatibilities with new software, or replacing the applications outright, can be expensive and complex.

Although IBM is now positioning its new z13 mainframe as the singular platform for cloud computing, there are other options for extending data to the web, which don’t always get their due. Part of the problem is that when it comes to extending mainframes into the ‘modern age,’ many IT groups are still working under preconceived, antiquated notions. I’d like to dispel a few of the more common modernization myths.

Myth #1: To start I have to “re-platform”

Hardware vendors are constantly bombarding companies with “upgrade” options – so much so that many now believe the only path to modernization is a full-scale replacement. Ironically, the advancements in mobile technology actually make that idea antiquated. Unlike legacy applications, mobile/web applications are not an extension of the backend. Instead, they are programmed separately and independently integrated with data sources via defined APIs. As long as they know how to talk to a mainframe, these ‘modern’ solutions don’t care how old it is. They will run optimally alongside their legacy counterparts for years.

Clearly, re-platforming is no longer necessary. Today, focusing on how to make technology serve the business – as opposed to serving itself – is the first step in any modernization program.

Myth #2: Open source and legacy don’t mix

Many assume that the proprietary nature of legacy systems will rebuff any attempts to introduce open source solutions into the IT environment as a whole. While that may be true when coding applications that run on the mainframe itself, it is not the case for modernized apps. Developers can easily use REST APIs to ensure connectivity in any open source environment. Because the mainframe will actually look like any other data source on a network, its information can then be accessed, updated or enriched via applications written in HTML5, Java, .NET, PHP as well as cloud-based and OpenStack solutions.

Myth #3: Web applications tied to mainframes will always be slow

Many believe that any application forced to communicate with a mainframe will automatically be subject to delays. Often, this is due to a tendency to over complicate the project with convoluted middleware solutions or an overuse of existing COBOL code snippets that are ‘close’ to the application’s proposed functionality. Response delays invariably occur in these situations.

Instead, developers should simplify the solution, code applications in Java and utilize APIs and connection pooling to communicate quickly and directly with the mainframe. Properly implemented, this method ensures millisecond responsiveness while actually reducing development timelines and project costs. Additionally, it allows developers to utilize IDEs, studios and other readily-available tools for quickly customizing layouts, integrating widgets and plug-ins and leveraging device-specific hardware features.

Myth #4: I can’t keep up with changing business requirements

Just because an IT infrastructure utilizes mainframe technology does not mean it cannot respond to evolving business requirements. Smart companies understand that agility doesn’t reside in the systems themselves. Instead, it’s more about balancing all available advances. For example, mainframes offer significant reliability and security advantages, however, native application development in COBOL can be extremely time consuming. Instead, today’s open source languages such as Java are extremely lightweight requiring 80% less code than its legacy counterpart. Simply ‘upgrading’ application development processes will enable IT groups to quickly deliver the innovative solutions employees and customers require, all without having to rebuild the entire datacenter each time a user requests a new feature.

Myth #5: It’s too expensive

Many IT groups are haunted by the hefty prices that they once paid for every new feature request or system integration project. Today, those expenses are no longer applicable. APIs simplify integration work and eliminate the need for costly middleware solutions. Open-source programming solutions that leverage mainframe data cut development time and infrastructure costs. In the end, this approach to mainframe modernization will actually lower Total Cost of Ownership – often by as much as 75%.

Unfortunately, all the talk about technology for the sake of technology can be extremely distracting to the modernization process. As IT managers sort through their options, it is important to focus on the true priority: creating business value from the information and systems at hand. Discussions should begin by asking, “what do our customers really need,” and “how can we boost employee productivity,” and not, “what should we replace next.”