Our education curriculum is designed to establish a comprehensive understanding of Unix architectures as they relate to important performance metrics. That foundation provides the basis to achieve validated measurements leading to meaningful performance analysis. Corrective action can then be based on fact, rather than speculation.

Classes are highly interactive and reference real-world case studies to illustrate how to maximize performance and ROI. Our lab sessions provide hands-on experience needed to successfully use WHAM's products in your own production environment.

Unix Architectures and Performance Optimization   

Prerequisites:	Solid understanding of Unix
Dates:	March 24-26
Location:	Austin, TX
Fee:	$1500
Status:	Open (12 seats available)

Training Agenda

1. Process characteristics 2. Thread characteristics 3. Priority calculation 4. Real memory management 5. Real memory control 6. Page space management 7. Address space management 8. Application resource measurements 9. Network architecture overview 10. Common data link interface 11. Network performance tuning 12. Filesystem implementation 13. JFS Filesystem

1. Process structure on Solaris 2. Threads on Solaris (User Level and LWP) Thread synchronization Mutex contention 3. Scheduling on Solaris Scheduling policies and terms Scheduling matters - a case study 4. Virtual memory management Address Translation (different kinds of addresses) Page space and virtual memory architecture 5. Address space management Heaps and How they are dealt with 6. Performance measurement architecture and tools on Solaris /proc interface and kstat interfaces ps, top, proctool 7. Performance profiling tools for processes gprof, truss, pstat 8. System accounting (SAR) 9. Performance measurement tools for the system vmstat, iostat, mpstat, netstat

1. Performance definitions 2. Scalability definitions 3. Measurement vs. modeling M/M/1 Queues and Little's Law 4. Accurate measurement Nyquist Rate Sampling Large effects of small errors 5. How to perform load testing 6. Netscape vs. Apache Case Study 7. Three Tier System Migration from Mainframe to Client-Server Scalability and Concurrency Limitations Through Prototypes How Performance was improved by a factor of 4 from Prototype Data 8. Same System Migrates to the Web with a Thin Client Interface Mutex Contention Case Study (Two Examples in One Application) Application Performance was improved by a factor of ten Cost was reduced by a factor of ten 9. A Computational Program Case Study How it was limited by Mutex Contention The Optimum Operational Configuration 10. Planning and Scheduling Application Case Studies How Poor IPC Messaging Limited the Performance One Application with multiple Performance Limitations The Impact of Improvements in Messaging Rate on Response Time

---

WebSphere & J2EE Hands-on Lab
WebSphere and J2EE are now in mainstream deployment within many corporations, especially in eCommerce applications. Whether you have already implemented these technologies and want to improve their performance, or you are in pre-implementation and want to optimize the size and configuration of your system, this is the ideal venue to learn more. This special lab session will provide hands-on experience with WHAM's performance optimization tools. WebSphere configuration and tuning considerations will be covered in detail, with particular attention given to known performance bottlenecks.

Prerequisites:	Unix Architectures and Performance Optimization
Dates:	March 27
Location:	Austin, TX
Fee:	$600
Status:	Open (6 seats available)

Training Agenda

1. WebSphere Performance Tuning Performance considerations Performance bottleneck identification Garbage collection 2. J2EE Performance considerations Performance bottleneck identification 3. FocalPoint Performance Tool Suite Introduction Resolver Refractor RangeFinder Silhouette Vista