ISG Provider Lens™ Mainframes - Services and Solutions - Mainframe Application Modernization and Transformation Services - Canada 2022
The cloud is pushing companies to integrate their mainframes.
Enterprises in Canada accelerated their digital transformation programs in response to the COVID-19 pandemic in 2020, with the public cloud infrastructure taking a central place in their digital strategy. Companies realized that on-premises infrastructures cannot match the resilience the cloud offers.
In the continuation of this strategy in 2021, mainframes have become a roadblock to digital transformation. Advisers, system integrators and software vendors have been reporting increasing demand for mainframe modernization, with three common patterns that repeat in every conversation: make mainframe data accessible through cloud applications and machine learning, improve agility with DevOps, and automate operations with self-service and provisioning by offering infrastructure as code.
This study identifies three possible scenarios. Some companies are in the early stages of digitalization and have priorities other than mainframe modernization. They prefer to leave their legacy applications untouched, running on outsourced data centers or mainframe-as-a-service (MFaaS) engagements. They have put modernization projects on hold until they complete other digital priorities.
A second group adds agility by modernizing their legacy COBOL applications with DevOps, APIs and microservices to integrate mainframe data with modern solutions and AI models.
A third group prefers to exit the mainframe and thoroughly embrace the cloud. To achieve this goal, these enterprises can use emulators, compilers and reengineering tools to automate the conversion of their mainframe applications to run in the cloud.
The mainframe operations market is mature and experimenting with organic growth. A small number of clients are migrating their mainframes from the incumbent provider, with most new workloads coming from on-premises data centers.
The mainframe as a service (MFaaS) market is a relatively new option to replace traditional mainframe outsourcing. In this model, the service provider offers shared mainframe infrastructure that enables clients to save on operating costs. MFaaS tries to match the public cloud experience, offering self-service over a standard service catalog, service dashboards and a pay-per-use commercial model. It also sees organic growth, with most new workloads coming from on-premises data centers. Clients can use MFaaS as an intermediary step until they are ready to modernize their applications.
Mainframe modernization focuses on agility and integration. It can be accomplished with COBOL applications modernization, upgrading older languages to COBOL and introducing a modern application development workbench, most likely with Eclipse and Microsoft Visual Studio and tools for automated testing, continuous integration, Java on z/OS, APIs, and DevOps. This market is growing at a modest pace. This study included a small survey sample, representing nine providers and approximately 1,000 enterprise clients. We found that only 7 percent use Java on z/ OS, 12 percent adopt mainframe DevOps and 15 percent use COBOL APIs.
Exiting the mainframe platform: Clients planning to exit their mainframes will find service providers referring to 7Rs — the seven possible strategies for legacy application modernization. These approaches are to rehost, refactor, reengineer, rearchitect, rewrite, replace or retire, which is an overly complex assessment model. Clients that are ready to exit their mainframes have long exhausted the replace or retire options. Rehost is the most preferred when enterprises want to exit the mainframe to x86 platforms (in the cloud). To exit mainframes and simplify the 7Rs, enterprises have only three methods, which are using cloud emulators, code compilers or rewriting applications. Understanding these methods is necessary to evaluate mainframe exit options, eliminate doubt and uncertainty, and manage risks.
The mainframe application modernization and transformation services market comprises providers that help clients exit their mainframes and migrate their applications to the cloud. They use proprietary and partner tools to automate the conversion to the cloud. AWS, Google and Microsoft have been pushing their partners to accelerate mainframe migrations. The cloud offers cost savings and improved performance when compared to mainframes. This service line is growing fast, with many providers reporting more than 20 percent revenue growth in the last year.
To prepare mainframe applications to run in the cloud, enterprises need automated tools to ensure quality, speed and predictable outcomes. Measuring the application portfolio lines of code (LOC) can give a rough estimate of application transformation costs. Refactoring with emulators and compilers can cost US$0.25 – US$2.30 per LOC and take anywhere from 2 to 36 months, with the cost of typical projects ranging from US$100,000 to US$25 million. Reengineering and rewriting applications can cost between US$0.50 and US$8.00 per LOC with project durations of 6 – 60 months and costs from US$100,000 to US$50 million.
Financing mainframe modernization is a challenge: Some clients that weigh the risk and benefits of application modernization are uncertain that the cost savings can compensate for the complexity and length of modernization projects. Vendors are investing in automating tools to act as a migration fabric. Clients should consider self-financing their application modernization with intermediary outcomes, procuring service providers that can plan phased transformations that take savings from one modernization phase to fund the next.
In the mainframe application modernization software market, this study qualified 17 vendors of mainframe migration tools. They partner with public cloud providers such as AWS, Google and Microsoft, and report more than 20 percent revenue growth in the last year. Each vendor delivers 15 – 30 projects per year, with an average project length of more than 18 months and large transformations taking up to five years. The market is slow in project execution, and there are relatively few projects. With the current pace, mainframe-to-cloud migrations may need more than 10 years to complete.
Access to the full report requires a subscription to ISG Research. Please contact us for subscription inquiries.