The AI Mainframe Optimizer is a Streamlit application designed to analyze mainframe job data (SMF-like logs) to identify optimization opportunities and provide recommendations for reducing MSU usage and costs.

• MSU Usage Analysis: Identify top MSU-consuming jobs and time periods
• Peak Detection: Highlight time windows with MSU overload
• Job Analysis: Detailed analysis of individual job performance
• Optimization Recommendations: Suggest job rescheduling to reduce peak MSU usage
• Savings Calculator: Estimate cost savings from implementing recommendations
• MSU Forecasting: Predict future MSU usage patterns using Prophet

• app.py: Main Streamlit application
• src/: Source code directory
  • data_processor.py: Data processing and analysis functions
  • forecast_engine.py: MSU usage forecasting using Prophet
  • visualization.py: Data visualization components
• utils/: Utility functions
  • helpers.py: Helper functions for formatting and summaries
• data/: Data directory containing sample data
• static/: Static assets directory

1. Clone the repository
2. Install dependencies:

   pip install streamlit pandas numpy plotly matplotlib prophet scikit-learn
   

1. Run the Streamlit application:

   streamlit run app.py
   

2. Upload SMF/CSV job data or use the provided sample data
3. Explore the different tabs to analyze MSU usage and get optimization recommendations

The application is deployed and accessible at: https://optizframe.streamlit.app/

The application includes sample data in CSV format with the following columns:

• Job Name: Name of the mainframe job
• MSU Used: MSU consumption for the job
• Start Time: Job start timestamp
• Duration (min): Job duration in minutes
• CPU Time (s): CPU time in seconds
• IO Count: Number of I/O operations
• Enqueue: Enqueue count
• Storage Used (MB): Storage usage in MB
• Job Class: Job classification

The application identifies optimization opportunities by:

1. Detecting peak MSU usage windows
2. Identifying jobs running during these peak periods
3. Finding alternative time slots with lower MSU usage
4. Calculating potential savings from rescheduling

• Integration with real-time mainframe monitoring systems
• Advanced machine learning for predictive job scheduling
• What-if simulator for testing different optimization scenarios
• Export functionality for reports and recommendations