OPTIMIZING A NEW MOBILE
CHANNEL WITH ESI
Better results with 50% less effort
“We would need twice the effort to produce similar results”
As soon as the new mobile app was launched a challenging goal was defined: rise users adoption from 10% to 60%. It was early detected, though, that performance issues were severely impacting the user experience. This new channel architecture transversed multiple layers of legacy systems before reaching the SOA services, which made troubleshooting extremely difficult. A multi-team task force was set in place to find bottlenecks and increase responsiveness.
How did ESI help?
Settle a common vision across multiple teams
Reveal solution architecture flaws
Find bottlenecks in SOA service
ESI solution impact:
0
0
0
ESI’s increased visibility led to the solution:
Each team had its own collection methods and metrics, so it was hard to reach a consensus. ESI versatility and ability to analyse data right from the level of the discrete events, instead of aggregates, was instrumental to dismiss some myths and to clarify some key points:
• All the interdependencies between services, their client and backend systems were directly extracted from ESIs just-in-time architecture diagrams and instantly made available to all members of the task force;
• Some user actions on the mobile app triggering more than one hundred service invocations;
• Some services were being repeatedly invoked, up to six-fold with the same inputs, in response to a single action at the presentation layer;
• One of the key services invoked by the app had an acceptable average response time of 60ms but 20% of its responses took over 300ms.
With the help of ESI, a mitigation plan was developed and a series of quick wins were put in place at different levels of the architecture stack. This quickly resulted in doubling the application performance.
