Many enterprises are either migrating applications from mainframes to SaaS/COTS or rewriting those applications entirely. But these projects take years to plan and implement, all while the cost of maintaining the mainframe piles up. Mainframe re-hosting can be an effective strategy to reduce those costs, creating self-funding for the mainframe elimination initiative.
Re-hosting is a relatively low risk approach that allows you to move your mainframe online and batch applications as-is to open systems. The re-hosting platform, which can be hosted on-premises or in the cloud, does not impact users since it provides equivalent services to the mainframe.
This option has become attractive in recent years as the cost of hardware has dropped significantly and support for legacy technologies within re-hosting platforms has matured. Commonly used legacy technologies like Cobol, JCL, PL1, Assembler, VSAM, IMS, and DB2 can now be hosted on these platforms. Re-hosting involves some one-time costs for new hardware and software licenses and the migration effort, but these can typically be recovered from operating savings in a year or two.
$12 million in savings
As an example, a financial services company’s mainframe system with 16M LOC, 9,000 batch jobs, and 2,000 user screens that was processing 3.5M transactions a day was successfully migrated to a UNIX environment. The primary technologies used in the source system were Cobol, JCL, Fortran, and IDMS.
The migration was completed in less than 18 months, resulting in approximately $12M in savings over a five-year period. Interestingly, the performance of the transactions in the new environment was much improved over the mainframe environment.
Annual costs before re-hosting (hardware and software): $3.6M
One time costs (platform & hardware licensing, migration services): $4.5M
New annual operating cost: $420,000
The majority of the migrations can be completed in less than a year, although more complex systems involving a multitude of interfaces and technologies take longer than that.
The benefits of adopting this approach become even more attractive when hosted on an IaaS cloud, which offers a utilization-based cost model for use of infrastructure services. Many of these systems have batch jobs that run weekly or monthly. In the cloud, the infrastructure services cost of executing those batch jobs is incurred only for the duration they run.
Taking a multi-phase approach
As you might expect, there is a small sub-set of technologies that are not supported by re-hosting platforms, so this solution cannot be used across the board. In companies using third-party applications or components (e.g., scheduler, output management, archiving) equivalent components must be found for open systems environments. Many of these have downstream integrations that have to be re-architected, so proper analysis of gaps in technology support and architectural change must be done to determine whether or not this is the right approach.
I recommend a multi-phased approach to the migration process:
Plan. Conduct an analysis to identify gaps in technology support and the need for new third-party solutions. Leverage this information to create a future-state architecture that includes infrastructure architecture. Develop upfront investment cost estimates to migrate the application that includes all license costs, migration costs, and costs for architectural and operational changes. Develop a business case for migration to help you determine whether or not the savings from the migration are significant enough to proceed.
Pilot. Migrate a slice of the mainframe application portfolio and do a limited release of the pilot application. It’s best to select applications that represent 5% to 10% of the code base of the mainframe and represent a cross section of technologies used in the mainframe. This phase will help you define strategies to mitigate unseen risks and enable you to calibrate your migration effort.
Migration. Create a factory model to migrate applications iteratively to the re-hosting platform.
Re-hosting is a low risk approach that’s often overlooked. It is an effective strategy to lower the operating costs of mainframe applications, and in cases where a decision to migrate to COTS/SaaS or rewrite the application has been made, it can be useful in creating self-funding for these initiatives. It should thus be part of the modernization toolbox of all enterprises.
Date de la publication : 2016-05-16