Performance

Engineering
Services

I. Definition of SPE: What is SPE and how should it be used?

• The software development environment, software lifecycle models, and integrated performance management
• Systematic SPE methods
• What should be done, when, and by whom, and how
• Introduction to adaptive SPE models for end-to-end system performance

II. Modeling Software Systems: What data do you need, and how do you construct software performance models?

• Data requirements
• Sources and techniques for gathering data
• Developing and evaluating software models
• Assessing scalability and sensitivity
• How to use the SPE models to evaluate software requirements and designs
• Computer hardware sizing, selection, and capacity planning
• Identification of performance bottlenecks and critical components
• Assessing scalability and sensitivity
• Software modeling workshop

III. General Principles for Designing Responsive Systems: What principles facilitate engineering performance into systems?

• The principles: fixing point, locality design, processing vs. frequency tradeoff, shared resources, parallel processing, spread the load, centering, instrumenting
• How to apply them to improve systems

IV. Software Performance Patterns and Antipatterns: What are some good performance-oriented solutions to common software problems, and what solutions repeatedly result in performance problems?

• The patterns: Fast Path, First Things First, Coupling, Alternate Routes, Flex Time, Batching, Slender Cyclic Functions
• The antipatterns: Excessive Dynamic Allocation, Circuitous Treasure Hunt, One Lane Bridge, Traffic Jam, Unbalanced Processing, Unnecessary Processing, The Ramp, More is Less, Sysiphus List Subset, The "god" Class, and newer antipatterns.

Predicting System Performance: How to represent contention among software processes.

• The system model of execution
• System modeling fundamentals
• How to model alternatives; how to evaluate scalability and sensitivity
• System Modeling Workshop

V. Software Performance Measurement: How to collect software performance data

• Concepts and Techniques
• Benchmark Planning
• Software Instrumentation
• Application Resource Measurement (ARM)

VI. Performance Assessment Techniques: How to collect the data you need and present SPE results.

• Strategies and techniques for data collection
• Conducting performance walkthroughs
• Controlling risks
• Presenting alternatives and recommendations

VII. Case Study: Lifecycle SPE for a Distributed System

• Performance walkthrough
• Teams construct and evaluate models,
• Teams present results and recommendations
• Summary and critique

VIII. Advanced Models of Complex Systems: How to represent complex systems and use advanced techniques to evaluate their performance.

• Modeling middleware: XML, SOAP, DCE, CORBA, MQSeries, etc.
• Service Oriented Architecture (SOA): Web Services and others
• Representing distributed interactions
• Component-based systems
• Commercial Off-The-Shelf (COTS) system integration
• Mainframe
• Batch
• Embedded systems
• Case Studies

IX. Review of Current SPE Tools: What tools are available to assist SPE efforts?

• Benchmarking
• Modeling
• Software development
• Tool inter-operability

X. Implementing Performance Engineering: How to apply this knowledge.

• Constructing effective systems
• Experiences with performance engineering
• Costs, benefits, risks
• Critical success factors
• Outsourced & purchased software
• SPE Technology Transfer
• SPE Business Case
• Establishing an SPE initiative

XI. Next Steps: How to Get Maximum Results from the Course

• What Are Our Strengths?
• What Have We Learned?
• What Is Our Vision of the Future?
• What Steps Can We Take to Achieve Our Vision?
• The 72-Hour Rule

Back to top

• Performance engineers who need to know how to model and analyze software designs; how to easily obtain the required data; and how to work effectively with design teams.
• Developers who need to know how to integrate performance-oriented design with software engineering practices; techniques for creating systems with good performance; and how performance models can be used to study design tradeoffs.
• Managers who need to understand the SPE process and how it can best be applied to their systems.
• Capacity planners or systems performance analysts who may support the software performance modeling activities or provide performance consulting services for software developers.
• Technology specialists, such as DBAs, network specialists, etc., who would like to create SPE models to predict the impact of future software systems on their environment.

Prior experience on software development projects is recommended; knowledge of software engineering practices is helpful, but not required. Experience in performance modeling is not required.