Case Study: Making the site faster.

Case Study:Making the site faster.

Case Study:  First American Bank

First American Bank is an Illinois-chartered, full-service bank with 50 Chicago area locations and $2.8 billion in assets.  The organization has an internal Web department staffed with developers and maintains a centralized website at firstambank.com.

The Need

The First American Bank Web team was concerned with the performance and response time of their newly rebuilt site.  In pre-launch testing, the site was noticeably sluggish.

The Approach

Imaginary assembled a technical team consisting of systems administration, programming and project management staff to audit the underlying technology of the site and provide recommendations to increase its performance.  The audit included server and core software configuration, database analysis and a review of software code.  In addition, Imaginary assessed the potential performance impact of each recommendation to assist First American Bank with a cost-benefit analysis.

Upon submission of the audit recommendations, First American Bank engaged Imaginary for implementation.

An Excerpt...

The following is an excerpt from the Technical Site Evaluation report produced by Imaginary Landscape for First American Bank:

"The NAS device holds the Web data itself, while all programs, such as Apache and Python, live on the virtual hard drive. One of the tests we conducted was to time how long it took to copy Web data. It took longer than 25 seconds to copy 25MB of data. We consider that much time to be unacceptable for a Web server. The server also exhibited sluggish performance when accessed via SSH, which may point towards server configuration issues."

The Outcome

In order to establish benchmarks for measuring performance improvement, Imaginary conducted a battery of tests before implementing the recommendations.  Each test was repeated numerous times with minimum, maximum and mean response time tracked.  Once the recommendations were successfully implemented, Imaginary conducted the same battery of tests for comparison. 

Results Summary

  • "Time to Connect" decreased an average of 43%
  • "Time to Transfer" data decreased an average of 90%