A Practical Approach to Mainframe Modernization
Mainframe systems still have a lot of supporters among IT professionals, and for good reason. A single IBM z16 mainframe can process as many as 19 billion credit card transactions every day. These systems support containerized applications, virtualization, AI and machine learning, API and cloud connectivity, and more. They’re also valued for their rock-solid reliability, boasting 99.999% uptime.
In light of this, modernization should not be viewed as a replacement strategy so much as an approach to unifying mainframe operations with today’s cloud platforms. To achieve that, IT leaders must understand the various methods and tools available to them, and how to design a manageable roadmap that balances costs, risks, and benefits.
Mainframe Modernization Patterns
Every organization’s IT landscape is unique, as are its business processes and requirements. Nevertheless, it can be helpful to segment the various functions performed by IT systems into a few broad categories.
Today’s cloud systems excel at high-volume data storage, powerful analytics, AI, and software & systems development. Mainframes have outstanding transaction processing capabilities, offering speed, security, scalability, and reliability. Modernization enables an organization to leverage the strengths of each platform, without making large-scale changes that can be both costly and disruptive.
There are three potential approaches to mainframe modernization:
Let’s examine each of these patterns in greater detail.
Data Replication
Many organizations begin their modernization journey to leverage powerful cloud analytics and decision support tools, including business intelligence, interactive queries, and real-time search. In the data replication pattern, information generally flows in one direction, from the mainframe to the cloud. Real-time or near real-time replication is most common.
领英推荐
To avoid burdening mainframe databases with constant I/O instructions and acknowledgments and prevent latency issues, best practices call for the use of event streaming platforms like Kafka, Amazon Kinesis, Rabbit MQ, or others. This facilitates a continuous streaming approach, allowing for extremely high throughput rates.
Regardless of whether or not a company decides to completely migrate off of the mainframe, data replication is a crucial first step that will be needed for any type of mainframe modernization strategy.
Cloud-based DevOps
Cloud-based DevOps facilitates the inclusion of mainframe data and mainframe development into the organization’s development efforts at large. This leads to more powerful and tightly aligned web applications, system APIs and functions, mobile apps and UIs, etc. The cloud-based DevOps pattern helps to eliminate friction between an organization’s increasingly interdependent cloud and mainframe development efforts.
Over time, the harmonization of mainframe and cloud development efforts will strengthen team cohesion and reduce both costs and timeframes for project delivery. Code quality and system interoperability will generally improve as well.
Mainframe Migration
Mainframe migration is a far more comprehensive approach involving the replacement of most or all mainframe applications and databases with cloud-native alternatives. This strategy generally involves significantly higher cost and risk because it requires large-scale changes to highly complex and interdependent systems.
Typically, a mainframe to cloud migration includes re-factoring code to a modern object-oriented language such as Java or C# and moving to a modern relational database. You must carefully consider various mainframe functions, including security, system logs, metadata, and COBOL copybooks when moving to the new cloud platform.
Given the complexity and scope of mainframe migration projects, most organizations seek help from outside experts, leveraging proven offerings such as the Mainframe Modernization Service offered by AWS.