As per the research done by Morgan Stanley, mobile internet will soon be overtaking fixed internet and we can see and feel the same trend now given that smartphones and tablets sales are overtaking PC sales,so all these changes compels us to go with a market demand.
So I thought let me share some of my thoughts based on my exposure to mobile world.I did poke around with WAP/Openwave emulator back in 2006/07 with both Silk Performer and LoadRunner.I do remember they have a protocol bundle for testing mobile traffic back then.I faintly remember it was more of an open source emulator support which these tools were supporting.Now the scenario is different, WAP is dead and now plain HTTP rules the mobile world.So we have some old rules which I believe still applies to mobile world and then there are also some rules which are totally new when we think in terms of Performance Engineering.
The strategy for designing for the mobile web site differs drastically compare to the desktop based web application given that applications works in the mobile context with limited amount of resources like screen size,network capacity,Hardware resources etc.The site needs to be lean in design and be accessible on the smallest mobile device with the exact functionality as offered on desktop based web applications.So the end to end strategy for testing mobile based application also differs compare to the desktop based web applications at times throwing unique challenges to the testing team.
It is also important to know and understand the capabilities of the various types of phones available in the market so that we have some idea as what is happening in the client side of the device.Some phones have Cursor based navigation, some phones have focus based navigations and some have touch and some multi touch based navigation.Low end devices normally have cursor or focus based navigation,so users using this focus or cursor based type of device will not normally prefer to have navigation depth of more than 3 links while browsing the site.The same user might browse a little longer in your site with either the tablet of smartphone.So it is fair to expect that page depth of the browsing in case of smartphones and tablets will be somewhat higher compare to Low end devices.This will help us in designing appropriate work load modelling for performance testing.
In this writing, I will try to summarize as what all things needs to be known and collected as part of requirement gathering for performance testing of the mobile web applications.
- Know your site along with its business domain and benefits it brings to users who are using your site.
- Know your users and how they use your application in mobile environment.
- Know the type of mobile device or handset your users uses to access the application.
- Know the network bandwidth your users uses in their device to access your application.
- Know the carrier your user uses to access your application.
- Know how long your users browses your site in the mobile space and collect the page depth of their browsing experiences.
- Know the type of the mobile browser your users use to access your application in the mobile environment.
- Know the user’s device’s screen size.
- Do not trust the web server logs to design the scenario as user behavior is not the same as compare to desktop users.If your business grows ,users multiply exponentially compare to desktop users given that users will be using your site on the move.
- There is very little or absolutely zero think time concept in Mobile space.Users are not going to wait to click the links unless they are verifying the some information before doing submit.
Sometimes it might happen that we do not have information on these points,in that I would suggest follow the mobile web standards and test taking those standards as your requirements.