Update applications with modern technologies previously unavailable to legacy systems. These may include containers for deploying a single application in multiple environments, machine learning and artificial intelligence (AI) for mining large data stores, and automation to first deploy applications and then monitor them over time.
The question of whether to migrate (rehost) or update (replatform) an application first is common. Rehosting typically results in immediate cost savings and compatibility with on-premises applications, Roth says. Replatforming, on the other hand, tends to speed up time to market.
Optimize applications on an ongoing basis to ensure the overall application modernization strategy meets the organization’s needs. “IT professionals know that as soon something gets to production, it becomes legacy and will need to be updated,” DeVries says.
To that end, he recommends a range of tools, services and best practices to guide the process.
Application Modernization and Rationalization in Healthcare
Not surprisingly, the unique nature of healthcare IT means organizations have additional considerations when it comes to application modernization.
It’s not unusual for healthcare applications to run code that’s decades old, whether it’s COBOL for administrative systems or MUMPS for clinical systems. Modernizing these applications is tricky, Fowkes says, as so many other systems depend on them. (That explains why they still run on mainframes.)
That said, there may be a way to apply modern technology such as generative AI to legacy applications. “What if you could tune a large language model on a 30-year-old code base, document what it does in plain English, optimize the code and potentially convert it to Java?” he says. “What cost can that take out, and how would that support the application modernization roadmap?”
In addition, many organizations need to confront the technical debt and application sprawl that stems from growth by mergers and acquisitions. While the situation has improved over the years, Fowkes explains, it’s still common for organizations to leave custom applications running in the background that meet one or two business needs but otherwise do no harm.
This is where application rationalization comes into play. It’s closely related to modernization, and it focuses on identifying inefficient or redundant applications. A cloud-native billing system may do what that custom application does — and it may work on all instances across the organization, not just the version running on a single machine. Or, as many health systems have learned recently, a single platform for virtual visits may do the same job as several implemented in haste in 2020.
“It’s important to understand the capabilities of new solutions and map them to existing software to allow intelligent decisions on what can or can’t be retired,” DeVries says. “Funding for new applications and technology can often be found by redirecting funds from software and solutions that can be retired or evolved.”
