Capacitas Logo

Capacity Planning and Performance Management
of Web Based Applications

Capacitas Training Centre,
London, United Kingdom.
4 Days

Learn how to:

  • Capacity plan an end-to end an intranet/internet based application
  • Understand the client-server and network components that make up a web based application
  • Define the performance characteristics of intranets, extranets and the Internet
  • Identify and apply a capacity planning methodology to a web based application
  • Build performance models to proactively capacity plan a web based application
  • Characterise the workload for a web based application
  • Capacity plan a web application using industry standard benchmarks
  • Gather the correct data in order to build a capacity model

Course Agenda

Day 1
9:00 - 9:30 Registration
09:30 - 10:00 A Capacity Planning Process for Web Services
  • The Capacitas Framework
  • Performance Data Collection
  • Workload Characterisation
  • Capacity and Performance Modelling
  • Model Validation
  • Capacity and Performance Prediction
  • Cost Modelling
  • The Capacity Plan
10:00 - 11:00 Performance Overview
  • Performance Defined
  • Why is Performance Important?
  • Client Performance
  • Web Server Performance
  • Internet / Intranet Performance
11:00 - 11:15 Morning Break
11:15 - 12:00 Web Infrastructures
  • Server Components
  • LAN Architectures
  • WAN Architectures
  • Web Service Protocols
  • TCP/IP Fundamentals
12:00 - 13:00 An Introduction to Performance of Web Based Systems
  • Breaking Down End-to-End Response Time
  • Queuing Terminology
  • Queuing Theory Primer
  • Upper Bounds on Performance
13:00 - 13:45 Lunch
13:45 - 14:45 An Introduction to Performance of Web Based Systems
  • A Simple Queuing Model
  • Understanding Arrival Rate Processes
  • Understanding Service Rate Patterns
  • System Level Models vs. Component Level Models
14:45 - 15:00 Afternoon Break
15:00 - 16:00 System Level Modelling of Web Based Systems
  • Modelling Systems with Fixed Sized Queues
  • Queue Sizing
  • Example: Performance of a Web Farm Load Balancer
  • Modelling Systems with Multiple Classes
  • Example: Performance of a Database Server with Multiple Transaction Types
 
Day 2
9:00 - 11:00 System Level Modelling of Web Based Systems
  • Modelling Systems with Priority Based Queuing
  • Example: Performance of a Corporate WAN Employing Prioritisation
  • Understanding Systems where Throughput Varies with Load
  • Example: Performance of a Web Farm
  • Example: Performance of a Web Farm with Throttling Employed
  • Example: A Capacity and Performance Model of a Corporate Intranet
11:00 - 11:15 Morning Break
11:15 - 13:00 Case Study: End to End Performance Modelling of a Web Based System
13:00 - 13:45 Lunch
13:45 - 14:45 Component Level Modelling of Web Based Systems
  • An Introduction to Queuing Networks (QN)
  • Open QN vs. Closed QN
  • Breaking Down Workload: Workload Classes
14:45 - 15:00 Afternoon Break
15:00 - 16:00 Component Level Modelling of Web Based Systems
  • Concurrency as a Performance Measure
  • Single Class Closed QN
  • Multi Class Closed QN
  • Performance of Multi Processor Servers
  • Example: Upgrading CPU/Disk Capacity of a Server to meet Service Levels
 
Day 3
9:00 - 10:00 System and Component Level Modelling Combined
  • Using Combined Models to Solve Complex Problems
  • Example: Modelling Concurrency Constraints at a Database Server
  • Background to Thread Pooling
  • Example: Modelling Thread Contention at a Server
10:00 - 11:00 Performance Measures for Modelling
  • Why Measure Performance?
  • Measuring Arrival Rates at the Web Server
  • Measuring Arrival Rates at the Network
  • Measuring Arrival Rates at the Database Server
  • Measuring Arrival Rates at the Payment Authorisation Platform
11:00 - 11:15 Morning Break
11:15 - 13:00 Case Study: Evaluating the Arrival Rate Patterns at a Web Server
13:00 - 13:45 Lunch
13:45 - 14:45 Performance Measures for Modelling
  • Evaluating Service Times Through Testing
  • Evaluating Service Times by Monitoring Production Systems
  • Using Benchmarks for Evaluating Service Times
14:45 - 15:00 Afternoon Break
15:00 - 16:00 Performance Measures for Modelling
  • Example: Evaluating Service Times for a Firewall
  • Example: Evaluating Service Times for IP Networks
  • Response Time Measurement Issues
 
Day 4
9:00 - 11:00 Case Study: A Testing Methodology for Measuring Service Times of an E-Commerce Platform
11:00 - 11:15 Morning Break
11:15 - 13:00 Bursty Workloads
  • What is a Bursty Workload?
  • Why do Bursty Workloads Occur?
  • Why are Bursty Workloads a Problem?
  • Measuring Burstiness
  • Modelling Bursty Workloads
  • Example: Sizing a Web Service to Handle Bursty Workloads
13:00 - 13:45 Lunch
13:45 - 14:45 Workload Characterisation of Web Services
  • The Relationship Between User, Workload and Performance Models
  • User Behaviour Modelling
  • Example: Building a User Behaviour Model
  • User Visit Modelling
  • Example: Building a User Visit Model
14:45 - 15:00 Afternoon Break
15:00 - 16:00 Workload Characterisation of Web Services
  • Summarisation Techniques for a User Model
  • User Modelling for New Services
  • Resource Level Workload Characterisation
  • Example: Resource Level Workload Characterisation for an E-Commerce Service

Course Brief

The growth of the internet/intranet, ease of deployment and maintenance are some of the reasons why web based applications are increasingly being introduced by ASPs and enterprises. The capacity planning analyst faces the challenge of having to plan the end-to-end system rather than individual platforms in isolation.

This course allows the analyst to plan and performance model a system comprising of web servers, database servers, LANs and WANs. The techniques discussed in the course can be used both to plan existing systems and size new systems.

Course Prerequisites

Course Audience

  • Capacity Planners
  • Web Developers
  • Network Planners
  • Network Engineering
  • Performance Analysts
  • Technical Architects
  • System Architects

How Your Organisation Will Benefit

Attending this course you will learn:
  • To use a capacity planning methodology for web based applications
  • How to ensure SLAs are not broken
  • How to perform end-to-end capacity planning of a client-server system
  • The performance characteristics of a web based application
  • How web and database servers impact application performance
  • How LAN and WANs impact application performance
  • To characterise the system workload
  • To build performance models to proactively capacity plan a web based application
  • How user behaviour affects system workload
  • Use benchmarks to aid the capacity planning process

Or consider an onsite event.

© Capacitas Ltd 2008 Privacy Policy Code of Professional Practice