Capacity planning is a strategic function used to predict IT resource requirements. It is a proactive extension of performance management, bringing order to chaos and predictability to IT management. Capacity planning and performance management together are sometimes referred to as capacity management.
This course provides an introduction to Capacity planning as a repeatable process. It's about understanding service levels and resource usage and aligning capacity requirements with business demands.
Capacity Planning training boot camp covers tactics for doing performance analysis, system sizing, performance measurements and capacity planning in today¹s challenging business and IT environment. Performance analysis and planning are often forgotten due to the project completion on time.
One of the major goals of capacity planning is to ensure that the business service level objectives are met. Identifying the relationship between Business Metrics of Interest (BMIs) and available system performance metrics is critical to accurate models and forecasts. BMIs represent the real world transactions that drive our business workload's resource consumption.
This course will define the characteristics of a good BMI and give techniques for identifying the best BMI choice from the many possibilities. We will discuss analytical tools, formulas and techniques used, and describe both theoretical ideals and real world cases. We will also show examples on how to use increasingly fuzzy BMIs to make great capacity plans.
Introduction to capacity planning
- What is capacity planning?
- Concepts and terminology
- Capacity planning and performance measurements
- Predict IT resource requirements
- A proactive extension of performance management
- Capacity management
- Network capacity planning
- Required components
- Administrative considerations
- System performance tuning
- Domain model review
- Process for network capacity planning
- Performance tuning & capacity planning
- Emphasis is on practical methods of measurement, simulation, and analytical modeling
Understand the IT resources needed to deliver acceptable service
- Accurate models and forecasts.
- Manage workloads
- Measure and manage performance and resource usage with business-oriented analysis
- Eliminate guesswork
- Isolate root cause of performance and response time bottlenecks rapidly
- Concentrate scarce resources
- Manage by exceptions from normal performance across complex environments via statistical analysis
- Identifying the relationship between Business Metrics of Interest (BMIs) and available system performance metrics
- BMIs represent the real world transactions that drive our business workload's resource consumption
The Capacity Planning Problem
- A Capacity Planning Methodology
- Performance Models
- Performance of Client-Server Systems
- Obtaining Input Parameters
- Model Calibration and Validation
- Performance of New Applications
- Supply Chain Optimization - 1:1-Type
- Supply Chain Optimization - N-Tier-Type
- Supply Chain Optimization - 1-Tier-Type
- Demand Capacity Planning
Capacity Planning Processes
- Why is capacity planning important?
- Foundation of performance
- Queuing, parallelism and multiprocessor systems
- IT management process maturity
- Find wasted capacity
- Model a single system or multi-tiered applications
- Overview of current procedures and processes
- Discussion of issues related to coordination of processes
- Performance analysis of an application
- Workload characterization Sizing using resource utilization profiles
- Application architecture-specific modeling
- Memory (buffers)
- Internal application scheduling and prioritization
- 3-tier architecture
- Users (client workstations, typically PCs, web users, etc.)
- Application Server
- Database Server
- The process to address the problem
- Understand where the data is located
- Collate and correlate data
- Build a model
Modeling Applications
- Modeling a pure 3-tier environment
- Measurements available for Database Server
- Separating relevant RDBMS instances
- Separating the 2-tier work
- Separating UNIX 2-tier users
- Using additional information from
- Application Servers
- Modeling a 2-tier environment
- Modeling UNIX 2-tier work
- Modeling PC 2-tier work
- Examples
- Sample resource breakdown
Capacity and Performance Management
- CPU
- Back plane or I/O
- Memory
- Interface and Pipe Sizes
- Queuing, Latency, and Jitter
- Application Characteristics
- Capacity and Performance Management Best Practices
- Service Level Management
- Minimizing the cost of meeting SLAs
- Reducing data center complexity
- Maximizing the utilization of resources
- Server consolidation
- Service level management
- Prediction of resource requirements
- A proactive extension of performance management
- Planning of hardware and software purchases
- Planning of upgrades and consolidate servers
- Capacity shortages or wasteful underutilization
- Analytical queuing modeling
- Capacity trending
- Historical patterns
- Linear trend line
- Linear growth path
- Predicts capacity requirements
- Infrastructure capacity planning
- Predict performance
- Determine the optimal configuration
- Components of response time
- Identify performance problems
- Manage performance with alarms
- Built-in performance guidelines
- Optimize IT infrastructure for peak performance
- Locate bottlenecks and pinpoint the cause
- Analyze real-time and historical performance
Modeling of Related Systems
- System structure
- The operational strategy
- Models
- Computer systems
- Communication networks
- The telephone network
- Data networks
- Local Area Networks (LAN)
- WANs and MANs
- Mobile communication systems
- Cellular systems
Probability Theory and Statistics
- Theory of Queuing
- Familiar queuing problems
- Characterizing a queue
- Basic metrics
- throughput, busy time, utilization, response time, load, service time
- Response time relationships for some simple queues
- Distribution functions
- Combination of random variables
- Time Interval Distributions
- Exponential distribution
- Steep distributions
- Flat distributions
- Cox distributions
- Other time distributions
- Observations of life{time distribution
Performance Management Processes
- Performance modeling of communication networks and computer architectures
- Performance characteristics and considerations
- Performance Quality Assurance
- QOS performance goal setting
- Network/Resource performance goal setting
- Subscriber service quality criteria
- QOS performance assessment
- Network/Resource performance assessment
- Resource performance assessment
- Data integrity check
- System and Network Management using SNMP and RMON
Performance Monitoring
- Performance monitoring policy
- Network performance monitoring event correlation and filtering
- Data aggregation and trending
- Data collection
- Traffic status
- Traffic performance monitoring
- Threshold crossing alert processing
- NE(s)/Resource trend analysis
- Performance monitoring data accumulation
- Detection, counting, storage and reporting
Performance Management Control
- Network traffic management policy
- Traffic control
- Traffic administration
- Performance administration
- Execution of traffic control
- Audit report
- Performance Analysis
Recommendations for performance improvement
- Exception threshold policy
- Traffic forecasting
- Customer service performance summary (excludes traffic)
- Customer traffic performance summary
- Traffic exception analysis
- Traffic capacity analysis
- Network performance characterization
- NE(s) performance characterization
- NE(s) traffic exception analysis
- NE(s) traffic capacity analysis
Performance Measurement Techniques
- Measurement
- Simulation
- Analytical modeling
- Performance Metrics
- Measurement Tools and Techniques
- Statistical Interpretations of Data
Analyzing Network Traffic
- Analysis process
- Network monitor
- Examining network packets
- Comparing network protocols
- Characterizing services
- Classifying network traffic
- Performance analysis techniques - a Review
- Performance data collection services
- Creation of predict models representing workloads
- Evaluation of the Predict model
- Verification and calibration of the predict model
- Predictive analysis of the effects of workload growth and system upgrades
- Graphs and reports to demonstrate modeling results
- Set response time and resource utilization guidelines and thresholds
- Review the impact of workload growth on the current resource utilization
- Evaluate the effects of various upgrades
Traffic Measurements
- Measuring principles and methods
- Continuous measurements
- Discrete measurements
- Theory of sampling
- Continuous measurements in an unlimited period
- Scanning method in an unlimited time period
- Numerical example
Optimizing Networks
- Network planning
- Network performance
- Optimization tools and techniques
- Optimizing client-to-server traffic
- Optimizing server-to-server traffic
- Predicting network traffic
- General guidelines
- Traffic prediction scenarios
Capacity Planning Applied
- Performance management vs. systems management
- Performance monitoring
- Performance analysis
- Performance prediction
- Server performance
- Application performance
- Application performance-benchmarks
- Performance management standards
- Industry performance management standards
Capacity planning and web services
- When Web Performance is a Problem
- Protocols and Interaction Models for Web Services
- Basic Performance Concepts
- Performance Issues of Web Services
- Planning the Capacity of Web Services
- Understanding and Characterizing the Workload
- Benchmarks and Performance Tests
- System-Level Performance Models
- Component-Level Performance Models
- Web Performance Modeling
- Availability of Web Services
- Workload Forecasting
- Measuring Performance
Capacity Planning Improvements
- Systematic method of consistently measuring servers
- Measuring application performance
- Lessons Learned
- Tools and Trend
- Tools for server and application performance management
- Case Studies
Labs and Hands-on Exercises
- Simulate projected traffic growth to identify necessary network upgrades and associated costs.
- Perform exhaustive failure analysis to determine risk areas, and plan redundant routes.
- Define service-level agreements (SLAs) for any metric, such as application response time.
- Design a cost-effective, optimized network that incorporates new technologies
- Optimize application deployments by right-sizing network resources, fine-tuning protocols, and configuring servers in advance.
