If the 1994 prediction were to come true, year 2022 would be 28th year of extinction of IBM Mainframe. But in contrast, mainframes are continuing their service and MIPS demand is growing. It continues to be a platform of choice for most of the large mainframe
shops. Then what is the issue? Why organizations must modernize the applications? What are the options available? In this blog, we talk about need for modernization, options, trends and how IBM mainframe is aligned to meet today’s and future demands.
Need of the Hour
The current mainframe applications are suffering from monolith-phobia, where conversation, function and data access layers are tightly coupled and make it difficult to distinguish the boundary. There is a lot of redundancy in data and code. The code complexity
and maintainability is a challenge. No clear documentation available, people who built those applications have already retired or on the verge of retirement. Also, most of the legacy skills are not available in the market. Arrival of newer and scalable technologies
have changed the business model. Customer experience is the main focus. Business operations have become global resulted in increased regulatory compliances and scope of risk analytics and fraud monitoring has become wider. The start-ups are new breed of competitors
now. These start-ups roll out new products and features in an unimaginable speed and slowly attracting the very customer base. The development methodologies have changed from waterfall to agile. To overcome the above challenges and grow, the need of the hour
is to modernize the monolith application for digitization, agility, speed to market and business scalability. So, modernization is not an option anymore. Then what are the options and trends?
Modernization Options
Whenever we speak about modernization, first thing comes to our mind is cloud. Cloud is being touted as an alternative to the mainframe. It has become a trend and it will continue for some more years. Also cloud offers modern set of tools and software for
development, DevOps and automated monitoring services. It is proven to be secured, resilient, guarantees availability and scalability. So, can organization, just shift the entire process on to cloud or re-write the code base into a new generation languages
and databases? How can one choose a right option?
There is no such thing as right option. It depends on the choice an organization makes based on its objectives, what it wants to address and amount of risk willing to take.
- Options like Re-Host and Re-Factor, are cost effective and less risky but does not address key SME and technology risk and faster time to market. One may speed up the delivery process by implementing DevOps but still coding and testing overhead will continue.
This option is fit for organizations with smaller Mainframe footprint, less than 2,000 MIPS and organizations with 5,000 MIPS or less can choose this as an interim option. - On the other hand, complete Re-Engineering or Re-Write is an ideal and futuristic option with application built on APIs, micro services based architecture deployed on a containerized platform. This option addresses key SME and technology risk. But it is
very high risk, requires huge investment, longer wait time to reap the benefits (minimum 3 years) and requires a long term commitment from both IT and business stakeholders. There are few or no use cases in industry for this option where entire system was
re-engineered on to cloud or on premise, however there are use cases where a subset of core differentiating functions were migrated to cloud or on-premise platform for agility and speed to market. - Hybrid and in-place modernization option provides opportunity to selectively modernize a subset of functions for agility and scalability. This is less risker compared to other options. This is also a cost effective and it does answer some of the key SME
and technology risks faced by organizations. a) Identify and expose core functions as APIs for consumption, such as customer on boarding, payments, loans processing etc.
b) Core data in real time or post process replicated to Cloud for AI, ML, Risk Analytics, AML, Fraud Detection etc.
c) Offload enquiry only function (CQRS), onto cloud as micro services deployed on a container based architecture
d) Participating in enterprise DevOps implementation and e) Use of impact analysis and documentation tools for quality deliverables and documentation to mitigate key SME risk.
Some references from industry on modernization in hybrid model, co-existence with Cloud thru selective modernization.
- A large investment bank implemented DevOps-CI/CD for a core application. Benefits are improved time-to-market, quality and reduced project cost.
- A large global banking and credit card processing customer has rewritten the core matching algorithm on to multi cloud platform. Benefits are microservices architected, configurable rules to improve agility.
- A large European exchange re-wrote its multi-asset class application on cloud using RedHat Openshift PaaS. Benefits are agility, speed to market, microservices design and Implemented DevOps CI/CD pipeline.
- A large US payments processing customer migrated one of their core processes onto Cloud platform. Code was refactored using automated tool first and then converted core portion to microservices. Benefits are agility in speed to market and de-risk of core
SME and technical skills. - A large global bank modernized its core banking functionalities by exposing them as APIs. Benefits are improved customer experience and faster service.
Sustaining and Investing into Mainframe Platform
IBM keeps upgrading its hardware and software to match the current trend and business growth.
- Offload development and testing on IBM cloud with Wazi as a Service (aaS).
- Expose core functional capabilities on system Z thru APIs to consume data using IBM z/OS Connect.
- Z Container Extensions, running Linux on Z docker containers inside z/OS – Any Linux software, available as a docker image can be deployed on Z.
- Tool sets support for DevOps and CI/CD pipeline implementation
Conclusion
Mainframes have been serving BFSI Customers well for many decades. They are robust and stable platforms supporting critical processing at BFSI firms. Since digital transformation is the key driver, instead of exiting the mainframe, organizations
should look at co-existence with cloud which is a less risky option. This option gives the best of the features from two worlds in terms of availability, resiliency and scalability, which is an added advantage and strength. Both platforms enable developers
to modernize applications without compromising security, scalability, and availability.