Planning an Effective Load Test

Planning an Effective Load Test (Design Phase):

Main activites in Design Phase can be of the following :

Gather Team

Define Goals

Outline Business Processes

Define User Profiles

Document Business Processes

Define Test Data Management Procedures

Outline Monitors

Document

Team

Business Analysts –Provide Usage Analysis of System Under Test (SUT)

Performance Engineers –Performs Business Process (BP) scripting and load scenarios executions

System Architect –Provides blueprint of the SUT

System Components SMEs –Provides Monitoring aspects of SUT components

Project Managers and Stakeholders –Keeping things running on track.

Goals :

Conceptual Goals

–What are we trying to achieve?

• Transaction Response Time of some Mission Critical operations

–Is the  transaction fast enough?

•SUT Environment Benchmarking

–Are servers, network, software, etc. compatible?

Define Goals:

The first step is to channel general application performance expectations into conceptual goals. Conceptual goals should outline all load objectives.

Some examples of conceptual goals and their relative priority might include the following:

Responsiveness of a “Search” function is a high priority example:

Are we able to get search results within a reasonable time?

Does “Update” transaction function well during heavy usage?

Gathering System usage Information:

By gathering system usage data you can:

 Decide on the business processes to record

 Isolate peak loads and peak load times

 Document user actions and valid input data for each business process.

From where can we gather the System usage Information?


a. First-hand use of the site.
b. Consultation with administrators to understand the nuts and bolts of the infrastructure.

c. Consultation with executives to understand the aspirations and expectations of the site.

d. Research of competing sites to understand the level of service in the marketplace.

Gathering system usage data will help you determine business processes you need to record, the expected response times, and the locations from which to monitor.

Business Processes

Mission Critical

–What BP malfunctioning will put you out of business?

Throughput

–How many users going through identified Mission Critical BPs?

Dynamic Content

–How much data each transaction process through the SUT?

Users

User Profiles

–What are the SUT users: level of experience with the system?

User Regions

–Where all SUT users located: AMERICAS, EMEA, APAC?

User Concurrency

–What is a pick of SUT usage per BP at given hour of the day (24h)?

Documenting BPs:

User action steps

-Should BP start with User Login?

–Can we navigate through the BP manually using these steps?

SUT navigation expected results

–What should we include in the Checkpoint once this step performed?

BP transactions

–What trigger in the BP will be measured for response time?

Think times

–How long user waits till the next step in the BP?

Parameters

–What static input data should be replaced with dynamic content?

–What type of parameters it should be (data file, date, random number etc.)?

–What parameter sequencing should be set (sequential, unique, random)?

Data Management

Test data needed

–What data will be needed (User Names & Passwords, SSN #s, SKU #s etc.)?

Procedures to generate data

–How are we going to generate data (querying database, data processing department, etc.)?

Generated data format

–What format data will be coming in (Flat file, CSV, etc.)?

Monitors

Investigate SUT environment

•Set monitor per each SUT component

•Add measure points gradually

–Set defaults measurements

–Add measurements as needed

–Consult with SUT component SMEs

Some points Explained in detail for the above:

Concurrency :

A concurrency is a set of users acting upon an application in a similar manner at the same time.

For examples if users are purchasing movie tickets online, we have the following Concurrency occurring on 3 levels as follows:

Application Level

– How many users are active on the system?

Business Process Level

– How many users are buying tickets?

Transaction Level

– How many users are buying tickets now?