June 17, 2025
Broadcom's Approach to Testing Mainframe Applications
7 min read
Written by: Uttam Anand
Key Takeaways
|
|
Mainframe application testing is essential to ensure the reliability, performance, and security of systems that support critical business functions across sectors such as banking, healthcare, insurance, and travel. These systems process large volumes of transactions and sensitive data, making rigorous testing vital to prevent downtime, data breaches, and system failures. Given their complexity and integration with both legacy and modern technologies, thorough testing is necessary to verify that updates, migrations, and new features do not introduce defects. Additionally, regulatory compliance requirements demand frequent testing to ensure data privacy and security are maintained.
Challenges in testing mainframe applications
Lack of expertise: Mainframe systems host legacy applications developed in older programming languages such as COBOL, Assembler, and PL/I. As experienced professionals with deep knowledge of these technologies retire, organizations face a growing skills gap. This shortage of expertise makes it increasingly difficult to test, maintain, and enhance these critical systems. For businesses that depend on mainframes to support core operations, this talent deficit poses a significant risk to stability, agility, and long-term innovation.
Limited availability of test environments: Mainframe systems are integral to business operations and are often required to run continuously, leaving little room for dedicated testing windows. This constraint significantly limits the ability to perform thorough validation and quality assurance. To overcome this challenge, organizations increasingly rely on advanced virtualization techniques to simulate real-world conditions, enabling effective testing without disrupting critical business processes.
High costs: Testing on live mainframe environments can be prohibitively expensive. Direct costs include hardware utilization, such as MIPS (million instructions per second) consumption, and increased energy usage. Indirect costs can arise from potential disruptions to critical business operations during testing activities. These expenses often limit the frequency and scope of testing, making it challenging for organizations to ensure quality without incurring significant financial impact.
Broadcom's approach to testing mainframe applications
The good news is that Broadcom DevTest, commonly known as Service Virtualization (SV), provides powerful capabilities to support mainframe application testing. SV enables organizations to virtualize mainframe transport protocols, making it possible to simulate complex interactions without needing access to live systems. This reduces dependency on limited test environments and helps teams perform testing earlier and more frequently.
SV currently supports key mainframe protocols, including:
- CICS Link
- CICS Transaction Gateway
- IMS Connect
- DRDA
Mainframe CICS communication architecture: An example using CICS Transaction Gateway
In this blog, I’ll use CICS Transaction Gateway (CTG) as an example to illustrate how SV can simplify and enhance testing within a mainframe environment.
The CICS communication architecture typically involves interactions between distributed applications and the mainframe using various protocols and interfaces. CTG serves as a bridge, enabling client applications—often running on distributed platforms—to securely invoke CICS programs hosted on the mainframe.
Mainframe CICS communication architecture
DevTest LPAR Agent overview
The DevTest LPAR Agent is a z/OS-based resident agent that operates as a started task, and it can be installed on any LPAR within the z/OS Sysplex. This agent acts as a central communication hub, facilitating seamless interaction between the DevTest Registry (and DevTest clients) and other DevTest agents running on z/OS. By coordinating these components, the LPAR Agent plays a critical role in enabling service virtualization and testing across mainframe systems.
By leveraging the DevTest LPAR Agent, DevTest can easily track the status of active z/OS CICS agents and access essential configuration details about those agents. Additionally, the LPAR Agent enables the grouping of multiple DevTest z/OS CICS Agents, which is particularly beneficial in environments where multiple LPAR Agents are deployed within a z/OS Sysplex. This capability enhances scalability and coordination across complex mainframe testing environments.
Modes of operation:
- Client mode: In client mode, the agent initiates the connection to the DevTest Registry. It is typically used in network environments where firewall restrictions prevent inbound traffic to the agent. By establishing outbound connections, this mode ensures communication is maintained without requiring changes to firewall rules.
- Server mode: In server mode, the LPAR Agent listens for and accepts incoming connections from the DevTest Registry. This mode is preferred when the network configuration allows for inbound traffic, as it enables easier communication from any mainframe bridge and simplifies integration in less restrictive environments.
DevTest CICS Agent overview
The DevTest CICS Agent is a resident agent designed for z/OS CICS environments, providing support for the VSE. It is implemented as a combination of the long-running task TKOA and a CICS exit.
The agent can be automatically activated through the CICS PLT (Program List Table) or manually managed via the following CICS transactions:
- TKOI for installation
- TKOR for removal
Agent configuration is straightforward, requiring only the specification of the DevTest LPAR Agent’s location to enable communication and integration.
Capturing mainframe interactions for seamless distributed system development
To facilitate the development and testing of distributed systems, it is essential to capture the expected responses from the mainframe when all back-end systems are fully operational. This process involves recording live interactions—both requests and responses—as they pass between the distributed system and the mainframe.
This approach, known as "on-the-wire" virtualization, enables teams to create a virtual model of the mainframe's behavior. Once recorded, this virtual model can be used to simulate the mainframe, allowing development and testing to continue independently of the actual mainframe environment. During virtualization, all data flows that would typically be routed to the mainframe are instead redirected to the virtual services layer, eliminating the need for a live mainframe connection. This results in faster, more flexible development cycles and reduces dependency on production systems.
Key benefits of on-the-wire virtualization
- Continuous and stable development: Ensures seamless development and testing of distributed applications without disruptions, allowing teams to work without dependency on the mainframe.
- Parallel development: Supports simultaneous development of distributed applications and mainframe systems, enhancing workflow efficiency and reducing project timelines.
- Cost efficiency: Significantly reduces mainframe-related expenses by eliminating the need for additional MIPS, disk space, software licenses, and other costly resources.
Conclusion
Testing mainframe applications is no longer a bottleneck for innovation—it’s an opportunity for transformation. As organizations face challenges such as limited expertise, high testing costs, and restricted access to mainframe environments, Broadcom’s DevTest solution offers a modern, flexible, and cost-effective approach. By virtualizing mainframe protocols through on-the-wire virtualization and leveraging powerful tools like the DevTest LPAR and CICS Agents, teams can simulate real-world scenarios, reduce their dependence on live systems, and accelerate both mainframe and distributed application development in parallel.
With Broadcom's approach, businesses can ensure the reliability, performance, and security of their mainframe systems while driving agility and innovation across their technology landscape.
Uttam Anand
Uttam Anand is a Client Service Consultant with comprehensive experience handling extensive engineering schematics and infrastructures. He has 12+ years of experience in DevOps and 16+ years of experience in the technology space. Uttam's expertise in implementing core DevOps Concepts, providing scalable and...