MIT’s Polaris Loads Web Pages Faster By Fetching Files More Effectively

A slow-loading webpage is a pain not just for end users but also for website owners. In order to get rid of this problem MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) and Harvard have developed a tool called Polaris that is claimed to reduce page-load times by 34 percent. Polaris framework determines the way how the page’s objects can be downloaded so that the overall website loads faster than usual. To understand how Polaris works you need to first understand how modern websites work. When you load web pages you are fetching elements like HTML files, JavaScript source code, and images. Once the browser starts receiving the elements it evaluates how the web page will be rendered. This process becomes considerably more time consuming when you realise that some elements depend on others to work. For example, a JavaScript has to execute in order to fetch images. The problem lies with HTML because it does not enable browser to know about such dependencies leading to follow a “as the crow flies” policy of cross-network trips and slowing down the page load.


Here is where Polaris comes in. It has been programmed to automatically track interactions between objects in a single page. Once it knows these interactions it generates a sort of dependency graph for the webpage. This enables the web browser to know which elements are dependent on which objects and fetch them in an optimised manner. One of the biggest plus points of the Polaris framework is that it is written in JavaScript which means that it can be employed on any website and no modifications are needed on the web browser. Website owners can implement this by running Polaris' dependency-tracking measurement platform on the site’s servers. The project Polaris was created by CSAIL PhD student Ravi Netravali who co-wrote the paper with graduate student Ameesh Goyal, Professor Hari Balakrishnan and Harvard professor James Mickens. Polaris will be showcased at USENIX Symposium on Networked Systems Design and Implementation (NSDI '16).

Source: #-Link-Snipped-# via #-Link-Snipped-#