Assessing and updating workhorse mainframes for the digital era isn’t a straight-through process. Here’s our guidance on how to move forward and reap rapid returns say T S Lakshmi Narasimhan and Niranjan Kulkarni, Cognizant Application Modernization Architects.

Part 1 of this series revealed how and why concerns around operating cost, skill sets and time required for new releases are prompting organizations to reevaluate the venerable mainframe platform against more digital-first compute environments. There are many reasons that businesses have decided to clutch long and hard to the mainframe and its many advantages.

These include a relatively low total cost of ownership for large enterprises with a significant number of users and transactional workloads; robustness, reliability and security built from many years of fine-tuning; and the complexity of business-critical applications.

Most of our midsize and large clients opt for this path, understanding that retention brings its own set of challenges amid today’s digital-first mandate. These include how to modernize monolithic, difficult-to-change legacy applications, usually with their data locked in silos, and increase speed of delivery. Companies must be willing to add open application programming interfaces (APIs) and integrated DevOps processes, and to selectively re-platform applications.

There are also skills shortages that will only worsen as proprietary applications continue to age. Irrespective of this, there is also a need to maintain legacy applications and continuously adapt them to digital shifts that put customer experience at the very center of the organization. Application teams must be willing to adopt cloud-native, hybrid application development models that will ease the skills shortage in the longer run.

Going forward, the mainframe platform will retain its value and long-term viability as these business modernize IT service delivery for the digital era. But achieving this requires enterprise-wide leadership and commitment to an adaptation process.

Tackling the problem

The updating process begins with an assessment of the company’s mainframe portfolio. This extensive phase comprises both quantitative (questionnaires, bug logs, etc.) and qualitative (stakeholder interviews, workshops, etc.) input. Mainframe applications are assessed for their potential alignment with digital; their flexibility and agility; and ways in which they are underperforming or failing to serve the business. Operating costs around mainframe infrastructure, maintenance and complexity are noted.

The assessment culminates in an analysis of the mainframe’s architecture and app integration; the company’s technology stack and infrastructure; the company’s code and data; nonfunctional requirements; pain areas and needs; and its digital/tools infrastructure.

With the analysis complete, needs and weaknesses are prioritized, culminating in a series of recommendations (from either a third party or the internal group spearheading the project), a roadmap to achieve goals, and cost estimates. There are two primary modes of updating:

1.   In-platform modernization.

This enables mainframe applications, data and processes to participate in the digital ecosystem. It is accomplished through services such as

  • “APIzation” (that is, the concept wherein multiple applications, both internal and external, can interact with and obtain data from one another based on an agreed-on set of inputs/outputs)
  • User interface modernization
  • “DevOpsification” that automates the software delivery lifecycle process, bringing in process agility
  • Technology stack standardization using accelerators like language conversion tools

2.   Right-platforming.

This entails re-platforming and re-architecting qualified workloads and applications, potentially into the cloud. Solutions include services such as re-hosting, rewriting, a hybrid of the two, and replacing. Of these options, the best fit is selected based on the assessment outcome. We have a set of our own accelerators, built over time through our work with clients, to accomplish seamless implementation of both in-platform and re-platform modernization.

Typically, large mainframe organizations will adopt a combination of these options, with the key factor being a given application’s strategic value to the business. The approach outlined above essentially involves transforming the traditional mainframe into a digital mainframe.

We advise clients to use this approach primarily to protect their investment in mainframe assets and reuse, both across the enterprise and with partners. Additional drivers could be simplifying the overall mainframe estate; eliminating disappearing technologies and skills (Assembly language, for example); and of course improving agility.

Off-platform modernization, by contrast — that is, migrating a workload from the mainframe to a distributed environment, whether on premise or in the cloud — is typically aimed at reducing operating expenses; aligning with a larger enterprise cloud strategy; eliminating risks related to the aging mainframe workforce; and moving to a cloud-based pay-by-use mode.

Modernizing to increase agility

In one engagement, we worked with the world’s largest package delivery company with a legacy IT portfolio (think mainframe, IBM’s Virtual Storage Access MethodSun’s J2EE, SQL Server, and so on) that was used to support online users for tracking or changing delivery instructions, with strict service-level agreements (SLAs) in place to propagate instructions to other systems.

With the influx of mobile channels for customers to convey and manage delivery instructions, this legacy process became cost-prohibitive to scale — and was deemed too rigid for the digital-first age.

The client sought to adapt the systems that its customers use to view the status of shipments and provide alternate delivery instructions. The goal: develop an IT strategy to modernize the app portfolio to newer, digital platforms — all built atop an architecture that provided greater agility in anticipation of the future’s inevitable (but nearly impossible to anticipate in detail) changes.

In the assessment phase, we workshopped and analyzed client data to gain an understanding of the existing product and application portfolio. In collaboration with the customer, we envisioned a future architecture that retained the legacy system’s capability and strength in operationalizing the delivery instructions well within SLAs, but decoupled systems of engagement (that is, customer-facing interfaces) and defined them on modern digital and agile platforms. Deliverables included a strategy and roadmap.

We helped the client team demonstrate that it can save $50 million in cumulative spend over three years, for a transformation investment of $12.5 million. The project was approved by senior management and is underway.

Moving to the cloud

Often, organizations seek to shift to the cloud for a complex blend of strategic, operational and cost reasons. This was the case for a leading telecom provider in the US. This client sought to move its existing ordering application from the mainframe to a Java microservices-based architecture hosted on AWS. This was one component of a larger strategic initiative to generally improve the IT infrastructure and move legacy applications to modernized platforms.

Beginning with the assessment phase discussed above, we gathered data on business and technical requirements. We also extracted the client’s business rules. This allowed the client to split the application portfolio into individual functionalities, then migrate them using a phased approach.

With data migration also a key concern for the client, we designed and implemented an approach to migrate data from DB2 to Apache Cassandra (an open-source distributed database) while ensuring that any data updates in Cassandra were available in DB2 until the applications were fully migrated.

Throughout the engagement, for each functionality that was moved off the mainframe, we designed and created new interfaces to ensure a smooth handshake between the new world and the prior mainframe functions. After the project is complete, the client expects the following results:

  1. A complexity reduction of approximately 30%
  2. Alignment with the overall enterprise cloud strategy
  3. Improved customer experience
  4. Better performance, scalability and flexibility
  5. A reduction in total cost of ownership

To learn more, visit the Digital Strategy section of our website or contact us.