Wednesday 9 March 2016

Researchers Develop Browser-agnostic System that Cuts Page Load Times by 34%

Waiting for pages to load when browsing the internet can be agonizing. It’s a concern not just for web users but also search giants such as Google, who are taking page load time into consideration when ranking websites. This led researchers at the Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Lab (CSAIL) together with their counterparts at Harvard to devise a system that makes page load times faster.

The teams from the two hallowed universities have come up with a system called Polaris, which overlaps downloading objects on a page so that the time taken to download the entire page is reduced.

Each time a browser has to fetch a piece of data by crossing to a mobile network, the process takes 100 milliseconds. The more complex a page, the higher is the number of trips required to fetch the required pieces of data. According to the report’s first author, Ravi Netravali, a PhD student, the complexity of a page can substantially add to its loading times.

To work out a solution to this issue, the researchers realized that they could reduce the time taken to load pages by reducing the number of trips to fetch data. The process becomes especially convoluted when the browser needs to fetch pieces known as ‘dependencies,’ which are essentially supporting objects required to a load an object on a webpage.

Polaris addresses this long-winded approach by giving the browser a list of objects that it needs to fetch beforehand – this enables the browser to calculate the fastest route in order to fetch the pre-decided list of objects. This is a departure from the conventional method in which the browser doesn’t know which objects it needs to fetch and in what order.

The team of researchers now hopes that Polaris will be integrated into web browsers sometime soon in the future so internet users can enjoy faster webpage loading times.

No comments:

Post a Comment