Having worked on a few new sites this year, we often had to cope with multiple merchants XML feeds, as is typical with travel sites. We started with the normal “serial” method which queries and collects the data from each in turn, i.e. when data is returned from Merchant 1, start the query on Merchant 2 and so on.

We then came up with the concept of “concurrent” queries, sending our search requests simultaneously and listening in the background for completions. This meant the user had to wait for as long as the slowest data returned, or when our set time limit had passed.

This idea isn’t new in other platforms and mediums where it is known as “threading”, and even using JavaScript as we had is fairly well known amongst developers. However, we found no clear cut explanations of how this works, so you can see HW Happy’s more detailed explanation here.

The practical benefits of this are pretty cool. If you have a product comparison site pulling live data, this could massively speed up your search times. It’s scalable, so whereas previously you might have limited the number of queries you sent due to time constraints, you can return the maximum information with no cost to the users time. There is no posting of data again and again to refresh the page, which gives the user a smoother interface. You can see one of our working examples on transfers.net an Airport Transfers comparison site.

Bookmark to: