Successful strategies for retiring mainframes

The mainframe may not be dead yet, but it’s got at least one processor in the grave.

Nearly 20 per cent of organizations, who responded to the Syncsort Inc. State of the Mainframe annual survey, plan to move completely off their mainframes in 2018. The major drivers that cause organizations to retire their mainframes include:

1. High hardware and operating system costs: Mainframes and their operating system are priced at a premium compared to x86 servers running Linux or Windows. For many organizations, this premium cost is significant and avoidable.
2. High software maintenance costs: Software development productivity is lower on mainframes. Therefore, making software changes is more time consuming and therefore more expensive.
3. High software development effort: Launching new services, typically dependent on software, takes too long. This elapsed time issue undermines responsiveness to business opportunities and competitor pressures.
4. Lack of access to best-in-class software packages: Many software vendors are no longer developing their software packages for mainframes. This software shortage imposes functionality shortcomings.
5. Rapidly aging skill pool: Software developers with mainframe skills are retiring and can be difficult to replace with individuals from the next generation. This resourcing issue creates business continuity risks.

In this environment, what are the effective strategies for retiring your mainframes? Many organizations are likely to pursue a blend of the strategies described below depending on the characteristics of each mainframe application in question.

Decommission or rationalize applications

Before you spend any money on migrating mainframe applications to a new computing environment, it’s useful to confirm that every application is still needed and being used. Over the years, organizations accumulate many informal applications that are later abandoned as the business changes and as end-users come and go.

It’s often possible to decommission or rationalize some applications by:

1. Simply turning them off.
2. Migrating any remaining end-users to use similar functionality in another existing application.

Rehost applications

One strategy is to simply rehost your existing portfolio of mainframe applications on x86 servers.

Rehosting means making only technical changes so that the applications will run, as before on the mainframe, under a different operating system and usually a different database management system (DBMS). In this strategy, the explicit goal is not to deliver any functionality enhancements, performance improvements or bug fixes.

The advantages of this strategy include:

1. Minimum cost and elapsed time.
2. Low project failure risk.
3. Little or no change to the work processes of your end-users and therefore little requirement for organizational change management (OCM) effort.

Implement software packages

Sometimes the best strategy is to recognize that your existing portfolio of mainframe applications is well past its best-before date. In this situation, the goal to retire mainframes becomes the trigger to replace and enhance all or most of your applications with a newer generation of software packages operating either on new x86 servers or in the cloud.

The advantages of this strategy include:

1. A greatly enhanced application portfolio that better meets business needs and can support the demands of your future growth.
2. Significant reduction in software maintenance effort associated with the predecessor custom applications.

The disadvantages of this strategy include:

1. Major changes in the business processes of your organization imposed by the software packages.
2. Major changes in the work processes of your end-users.
3. Significant OCM effort and cost.

Migrate applications to the cloud

Sometimes the best strategy is to use the goal of retiring mainframes as a trigger to migrate all or most of your mainframe applications to the cloud that uses x86 servers. This strategy is like the rehost applications strategy discussed above.

The benefits of the cloud and therefore the advantages of this strategy include:

1. No capital-expenditure for a computing environment. You are not buying servers or renovating a data center to retire mainframes. You’re paying for computing capacity as you go.
2. Access to instantly varying computing capacity. You are no longer buying for your peak capacity requirement.
3. High availability applications. You are no longer staffing and managing to a Service Level Agreement (SLA). The vendor now takes on responsibility for meeting the SLA.
4. Managed disaster recovery. You are outsourcing this service to your vendor.
5. Automatic software updates. You are no longer updating and patching your servers.
6. Managed security. You are outsourcing this service to your vendor.

Redevelop applications

Sometimes the best strategy is to recognize that your existing portfolio of mainframe applications is well past its best-before date. In this situation, the goal to retire mainframes becomes the trigger to replace and enhance all or most of your custom applications to operating either on new x86 servers or in the cloud.

The advantages of this strategy include:

1. A greatly enhanced application portfolio that better meets business needs and can support the demands of your future growth.
2. Total control over the direction of your application functionality and its pace.

The disadvantages of this strategy include:

1. The cost and elapsed time associated with custom software development.
2. The risk of project failure.
3. Major changes, hopefully improvements, to the business processes of your organization.
4. Major changes in the work processes of your end-users.
5. Significant OCM effort and cost.

Address informal applications

One issue, that’s common to all strategies for retiring mainframes, is how to handle informal applications that were created by pro-active, energetic end-users to extract data from your mainframe applications to support a business process with useful reports and charts. Every one of these informal applications will require some software development to function within the new application portfolio. These informal applications typically use tools like Microsoft Access, Microsoft Excel, Tibco Spotfire or Tableau.

One approach, to pre-empt noisy complaints about retiring mainframes from these end-users, is to provide a little coaching to help them make the required changes. A disastrous approach is to have the IS support group start to look after these informal applications because that will consume huge effort for little business benefit.

What issues have you encountered as you contemplate retiring mainframes? Let us know in the comments below.