Strategy of device selection

Owned by Maeva Rondeau
Jul 26, 2024
2 min read

Give the fragmentation of Android’s smartphones, it is necessary to optimize your device choice. In fact, it is impossible to measure energy and resources consumption for every device on the market.

Several strategies are possible for choosing the right smartphones to include in a test campaign. First of all, we need to define some validation objectives :

  • Should we focus on a reference device ?

  • Should we follow consumption on a representative range of users ?

  • Should we use a less powerful device to detect performance problems ?

Devices classification

A first sort can be made by using smartphones ranges :

 

Low end

Mid range

High end

 

Low end

Mid range

High end

Used technologies

Older-generation technologies :

  • single core

  • <1Go RAM memory

 

Latest technological developments :

  • quad or octo core

  • >4Go RAM memory

Purpose

  • Reveals performance problems

  • Verify if the application is usable by all users and avoid to exclude certain people.

  • Validate the largest number of user

  • This type of devices are usually the most common.

  • Verify the smoothness

  • Impact of of latest technology components

Range choice

To reduce tests possibility, it may be possible to remove some ranges :

  • For an application intended of a targeted audience, for example for a specific professional use or panel of customers, we will only test on the appropriate range.

  • If performance is the focus of measurement, a low-end device will be enough to validate higher-end devices.

  • For maximum coverage of a wide variety of users, a test at the limits with a low-end and a high-end device will be relevant.

A precise choice will always be better : an analysis of the smartphones most commonly used with the application will enable to prepare measures conditions as close as possible to real-life.

Strategy application

This strategy must be coupled with continuous integration.

It may be necessary to continually monitor the most critical device and measure occasionally an other device from a different range.

We can detect potential problems with this method and also confirm our critical device choice and the relevance of our device choice on the main continuous monitoring.