BUSINESS NEED :

Fashion & Beauty brand was looking to develop their application with greater efficiency, so that maximum number of user access can be handled smoothly. There should not be any error, resulting in freeze & crash of the mobile app.

RESULT :

Andolasoft proposed to go for Amazon Web Services (AWS) and its Elastic Compute Cloud (EC2) service. A few hours of research showed that AWS is certainly more reliable, cost effective and used as backend of Heroku, providing a greater control over installing the software

Andolasoft helped PrettyInMyPocket (PRIMP) to migrate their production server infrastructure from Heroku to Amazon Web Services (AWS) and helped reducing the monthly expenses

Overview

This case study narrates how Andolasoft helped PrettyInMyPocket (PRIMP) to migrate their production server infrastructure from Heroku to Amazon Web Services (AWS) and helped reducing the monthly expenses.

Pretty in my Pocket (PRIMP) is a powerful way to shop for beauty products online, with access to product ratings, reviews and a real community.

PRIMP already had their application deployed on Heroku server, running with two standard-2x dynos in addition to configuration for auto-scaling and a separate plug-in for web solr. The database was running with Heroku’s Postgres add-on.

Due to increase in user base, number of hits to the server had gone up, making the existing dynos in Heroku unable to handle the load.

Most of the time the application suffered from R14, R15 (Full Memory), H12 and H13 (Connection Timeout) error, resulting in freeze & crash of the mobile app.

There was no separate server to run the background jobs, consuming high CPU power.

To handle the situation the dyno size had been increased, which further increased the monthly budget to more than $1000.

There was a huge difference in the code & configuration for production and staging environments.

Moreover, the server infra was totally managed by Heroku with no SSH option.

The Amazon EC2 virtual server instances come with a wide range of CPU and memory configurations. In order to run the application with greater efficiency, we had to choose the instance depending upon the type of tasks it will handle. Also we had to keep the budget in mind.

Therefore, we chose different type of EC2 instances for serving different purposes. Deployment to AWS was not as easy as Heroku, but with Elastic Beanstalk’s easy configuration it wasn’t tough.

In order to overcome these issues, Andolasoft proposed to go for Amazon Web Services (AWS) & its Elastic Compute Cloud (EC2) service.

A few hours of research showed that AWS is certainly more reliable, cost effective and used as backend of Heroku, providing a greater control over installing the software you need; Unlike in Heroku.

For Seamless Migrations…

Related Case Studies