Application performance analysis in industry
Why is application performance measurement necessary but inadequate? Because although the measurement informs us, it does not enable us to interpret or anticipate. The ultimate aim of an information system is to aid growth by ensuring the right decisions are made at the right time.
Imagine your trucks are stuck at the garage because they have not been authorized to leave. Imagine your chain of stores with each of its hundreds of points of sale wasting 10 seconds per transaction. Imagine your Australian employees struggling to receive the data required to make production progress because your application infrastructure (ERP, CRM. etc.) is based in Europe. Businesses depend on critical applications; a drop in performance or unavailability puts them in a vulnerable situation.
Examples of application performance analysis
Let us start with a simple example in the luxury goods sector. Like a physical store, the performance of an e-commerce site requires numerous applications to operate correctly: from in-store revenue management and the electronic payment system to the back office which manages the supply of raw materials. Imagine that following a successful marketing campaign your stores experience a peak in activity, perhaps resulting in the payment platform becoming overloaded. A good information system can identify the network fault and justify the necessary investments.
Now let us take the example of an e-commerce site. A mature consumer who is unable to pay for purchases will lose confidence (and patience) and go and buy the product elsewhere. The application performance analysis will explain the momentary drop in revenues: the consumer was simply unable to pay!
Application performance analysis provides confidence and clarity by identifying weaknesses and useful areas for investment.
Moving on to the energy industry, the large number of sites abroad means that there is a large volume of data to process and multiple fluctuations relating to the network, as well as numerous material and non-material exchanges and therefore an infinite number of operations relating to the information system. How can we make sure that everything is working well on a global scale when there are networks of radically varying quality in the different geographic areas?
Let us take the example of a multinational with a site in Tanzania, one in Johannesburg and one in New York. How can we assess application performance?
It is important to combine two types of measurement:
- – a network time measurement (time taken to forward a data packet to a data center in milliseconds),
- – an application measurement (execution of an application script which is used to assess the user experience).
We start by testing network performance (in milliseconds) then refine the analysis by measuring business applications (in seconds or minutes) on flagship sites using test units which run user scenarios. These robots may cost money but it is here that application performance analysis comes into its own with minimum cost for maximum results. But more specifically, how can the user experience be assessed on 3 sites with 2 robots?
Application performance analysis: the art of deduction
- Choose two flagship sites: Johannesburg and New York.
- Measure network quality: 200 milliseconds in Johannesburg, 50 milliseconds in New York.
- Execute the business script: 2 min in Johannesburg, 1 min in New York.
- In Tanzania, we only measure the network quality: 180 milliseconds (remember that it is 200 milliseconds in Johannesburg!)
- We can therefore deduce that in Tanzania, the employee executes the application scenario in (approximately) 2 minutes.
Using efficient applications guarantees the chance of long-term success (increased revenues, better information system between subsidiaries). Investing in effective application performance analysis provides you with a decision-making aid and therefore development tool. This is the added value of the managed service offered by Maltem Insight Performance.