Since the early days at Altersis Performance we were aware of the complexity of the performance problematic within the context of distributed and composite applications.
The complication lies in the number of architecture tiers (entry, presentation, business logic and data), of runtime platform layers (OS, JVM, application server, frameworks, and applications), the number of custom and reusable code components, and the complexity of the business transactions that are behind.
Altersis Performance has elaborated the first measurement-based, adaptive, and life cycle performance engineering methodology called PEP: Performance Engineering Process. Our approach targets performance assurance in the highly complex context of composite applications and focuses on the features of state-of-the-art architectures, key technologies as well as leveraging existing approaches and concepts.
PEP consists of the application of a disciplined, systematic and measurable approach to define, evaluate, optimize and monitor application performance.
PEP: The Patterns
This performance engineering methodology, successfully used and implemented in hundreds of projects over the years, is based on three key patterns:
- Scenario-Driven Requirements Modeling (SDRM) to enable performance requirements engineering from development through QA up to production monitoring
- Application-Centric Measurement and Monitoring (ACMM) as the central piece for designing, implementing, end-to-end and life cycle instrumentation as well as monitoring solutions.
- Life Cycle Integration and Continuity (LCIC) in form of workflows and templates enables a smooth implementation of the performance engineering activities into the existing processes and workflows.
These patterns are applied to all three major phases (Build / Pre-Production / Production) of the application lifecycle with dedicated activities, environments and staging rules.