Customer stories

CDN testing and migration halves workload and saves $70K

2 min read

The challenge

The client needed to validate the migration of more than 50,000 web pages across 250 domains to a new content delivery platform.

The scale and sensitivity of the content required precise testing. The solution also needed automated reporting and reusable components for future migrations.

The solution

Beyondsoft developed an automated tool to compare:

  • Link redirects and static content.
  • HAR files and MD5 values.
  • Screenshots captured before and after migration.

The tool generates HTML reports with comparison scores for each page. For mobile application pages, it compares pre-login API calls.

Separate Docker containers test each pair of existing and new domains, supporting the large testing workload.

Additionally, to comply with the customer’s acceptance criteria, Beyondsoft delivered test strategy documentation, insights around the reusability level of the automated tool, crawler scripts and documentation around their accuracy, and proof of the tool’s ability to test multiple languages.

Implementing the CDN testing and migration solution

The tool uses a modular workflow that can be reused for future migrations:

  1. Receive the new domains.
  2. Check whether screenshots and HAR files already exist.
  3. Run a crawler to collect missing files.
  4. Compare the existing and migrated content.
  5. Perform multiple testing passes to improve accuracy.
  6. Generate a report with scores for each test type.

The biggest implementation challenge was ensuring the tool could handle multiple aspects of testing without failure. Today this tool can be used for any future CDN migrations. The tool is extremely versatile and can be adjusted to the specific needs of any client.

How the testing solution works

CDN testing and migration architecture

The results

Thanks to the reusable testing and migration solution developed by Beyondsoft, our customer realized the following benefits:

  • Decreased manual workload by 50%.
  • Improved process efficiency by 75%.
  • Reduced the time required for testing from two months to just five weeks.
  • Lowered the cost from $250K to $180K.

Technologies used

Java, Python 3.8, Firefox 81.X, and Geckodriver v0.28.0