In-app Response Time Measurements

By Author: Sampath Prakash
Total Articles: 2
Comment this article

What can be defined clearly can be measured and what can be measured has a hope of getting controlled and managed – that

is so true for end-user response time for applications. In-App response time measurement is an excellent approach for

accurately measuring true end-user response time.

When an application is built, there has to be awareness of the performance of at least the important transactions (if not all

transactions). The developer has to provision the code for timing the beginning and completion of a transaction and storing

the value of the resulting response time. This information can be further sent to a cloud-based system for storing and

analysis.

An example is when a user launches an App such as Netflix on a tablet and receives the full home page or when a user is

checking the description of a movie of interest. The App can measure this time interval and Netflix can aggregate this

information for many such transactions from many users and analyze the data for performance improvement.

Providing In-App timers and measurements ...
... is not a new idea. All the popular browsers (IE, Mozilla, and Chrome) implement

this. The browsers can measure the page download time and by means of java scripts they can transmit this information to a

cloud server for aggregation and analysis. For example, Opnet’s ARX Browser Matrix Software-as-a-Service is based on this

feature and helps track response times or page download times for websites.

Another open source project that exploits browser In-App measurements for page download times is www.webpagetest.org.

Any website URL can be tested from selected global locations and very useful results are delivered in a couple of minutes or

even seconds. In addition to the page download times the site provides valuable information about the composition/size of the

objects of the web page and the sequence in which they are delivered.

Web-based applications are most common with the current installed base PC/laptop devices and hence we can take advantage

of these browser-based In-App measurements for Application Performance Management (APM). However the recent trends

show proliferation of mobile devices -smartphones and tablets. There are many non-browser based applications on these

devices and to implement In-App measurements within these Apps could be a big challenge. These are typically low cost

Apps and to implement In-App measurements could easily increase development cost by significant amount.

The Apps on the mobile devices are the ones that encounter low bandwidth, high latency, or high packet loss situations

resulting in poor App performance. They are the ones which could mostly benefit from these In-App measurements but can

least afford.

In-App measurements require a significant change to the current software development practices and will take time. What we

expect is that for those applications with mission critical needs (e.g., healthcare) or financial trading, or e-commerce

transactions In-App measurements will first appear for web and non-web based client server applications. The Open Group

forum’s Application Response Measurement effort is a welcome step in the right direction.

Next time when you do a Google search make a note of the how long it took – this is displayed prominently in the very first

line. That is what we want for all App transactions!

Apsera Tech has years of experience in WAN optimization, networking and application performance management. Apsera Tech’s APM consulting services have helped Fortune 1000 companies in diverse industries in planning/resolving critical business application performance issues like slow application response time, WAN optimization and WAN acceleration. For more details, please visit our website at www.apseratech.com.