The applications are responsible for house data and price information, respectively, which are retrieved from two different sources. One of them uses message queues to collect price data, and the other listens to a series of web services that return XML. Both applications transform and cleanse the data and return JSON to the services that need this data for further work.

The price information is distributed across message queues and covers a wide range of transactions that need to be managed in terms of sequence and timing, which has placed high demands on the application's performance in terms of processing time per message. Therefore, the applications have undergone extensive testing using jMeter, which simulates a transfer from start to finish.

The house information is not performance-heavy, as the application runs as a batch operation once a day. The application connects to a web service and, after token authentication, retrieves a large XML document, which is processed and further specified with data from two other external services. The cleaned data is sent in Zip-compressed JSON format to Drupal, which then updates each content element through a specially developed module that the provider has developed in PHP.