Contact Us
    AIOps

    The Evolution of Mainframe Operations: From ITOps to the Promise of GenAI

    August 14, 2024

     

    In the ever-changing landscape of mainframe operations, one constant remains: the need for efficient, reliable system monitoring. Today, we'll follow Ashton, an IT operator, as he navigates the transformative journey from traditional ITOps to the cutting-edge world of Generative AI-assisted operations. 

    The ITOps Era: Manual Monitoring and Firefighting 

    It's 2010, and Ashton starts his day by logging into multiple consoles to monitor CICS (Customer Information Control System) performance. His eyes dart between screens, searching for anomalies in transaction response times, resource utilization, and error rates. Suddenly, an alert flashes: 

    CRITICAL: CICS region PROD1-Transaction ABCD response
    time exceeds threshold

    Ashton’s heart races as he begins the time-consuming process of diagnosing the issue. He manually combs through logs, checks system resources, and correlates events from different subsystems. After hours of investigation and several escalations, he identifies a poorly optimized database query as the root cause. 

    The AIOps Revolution: Intelligent Insights 

    In this era, Ashton benefited from capabilities provided by Broadcom’s WatchTower Platformfor Mainframe AIOps. His workstation now features a centralized dashboard that ingests data from multiple mainframe subsystems. As he settles in for his shift, the AI-powered system has already detected an emerging issue: 

    ALERT: Anomaly detected in CICS region PROD2 - Click here to view visualizations of anomalous metrics

    The AIOps platform assists Ashton with: 

    • Root cause analysis
    • Proactive and predictive Insights on potential impacts
    • Recommended remediation steps 

    Instead of spending hours diagnosing the problem, Ashton quickly identifies the root cause and addresses it effectively, preventing a major outage. He marvels at how AIOps has transformed his role from reactive firefighting to efficient root cause analysis. 

    The GenAI Era: A Glimpse into the Future 

    As we look ahead to the near future, let's imagine how Ashton's workday might transform with the advent of Generative AI: 

    As the time progresses, Ashton sits down at his desk, greeted by an advanced AI assistant: 

    "Good morning, Ashton. I've analyzed the overnight CICS performance data and have some insights to share."

    Ashton engages in a natural language conversation with the AI:
    Ashton: "Thanks. Any concerns I should be aware of?"

    GenAI: "Yes, I've detected an unusual pattern in transaction EFGH. While it's not critical yet, historical data suggests it could lead to a bottleneck in PROD3 within the next 4 hours."

    The Generative AI assistant instantly generates a detailed report, complete with visualizations of the transaction flow, system resource utilization, and code snippets highlighting potential optimization areas. It also proposes possible solutions, ranking them based on effectiveness and implementation complexity.

    Ashton reviews the suggestions and decides to implement the top-ranked solution. He asks the Generative AI assistant to draft the change request and simulation results, which he quickly reviews and approves.

    Within minutes, the change is implemented, tested in a simulated environment, and deployed to production. Ashton monitors the results in real-time, watching as the potential bottleneck dissipates before it ever becomes a problem.

    The Future is Within Reach

    While this scenario may seem futuristic, the technology to make it a reality is already in development within the generative AI community. At Broadcom's Mainframe Software Division, we're at the forefront of this AI-driven revolution. We're actively researching and working with open-source Large Language Models (LLMs), fine-tuning them with proprietary mainframe data to create powerful, specialized AI assistants for mainframe operations.

    Our approach offers several key advantages:

    1. Data Privacy: Unlike cloud-based solutions that require sharing sensitive data, our on-premises approach allows customers to leverage advanced AI capabilities while keeping their data secure within their own infrastructure.
    2. Customization: By fine-tuning open-source LLMs with customer-specific data, we can create AI assistants that truly understand the unique complexities of each mainframe environment.
    3. Continuous Improvement: As the AI learns from ongoing interactions and new data, its capabilities will continuously evolve to meet the changing needs of mainframe operations.

    Partnership Opportunities

    We believe that the future of mainframe operations lies in collaborative innovation. That's why we're inviting mainframe professionals, data scientists, and organizations to partner with us in this exciting journey. By working together, we can:

    • Develop use cases that address real-world mainframe operation challenges
    • Refine and expand the capabilities of our AI models
    • Create best practices for integrating GenAI into mainframe operations
    • Ensure that the solutions we develop truly meet the needs of the mainframe community

    Join Us in Shaping the Future

    As we stand on the brink of this transformative era, we invite you to join us in exploring the possibilities of Generative AI in mainframe operations. Whether you're an IT professional like Ashton, a mainframe architect, or a business leader looking to drive innovation, there's a place for you in this revolution.

    By partnering with Broadcom's Mainframe Software Division, you'll gain early access to cutting-edge AI technologies, contribute to the development of industry-leading solutions, and position your organization at the forefront of mainframe innovation.

    The future of mainframe operations is not just about technology – it's about the synergy between human expertise and AI capabilities. Together, we can drive unprecedented levels of efficiency, reliability, and innovation in mainframe operations.

    Are you ready to be part of shaping the next era of mainframe operations excellence? You can reach me at mohit.dadu@broadcom.com to learn more about our Generative AI initiatives in mainframe operations and discover how you can get involved.

    Tag(s): AIOps, Mainframe