So as you are probably aware Google’s latest search innovation google instant that gives predictive search results and has been launched to much fanfare today.  Google says it will speed up search and get people their results faster. This is done by using an AJAX technique ( technically as it doesn’t use XML but JSON in it’s response I guess it is AJAJ but that doesn’t sound as good) with a HTTP request being sent after every letter you type which returns an updated set of results.

Now from a usability point of view this a great, however it comes at an expense as each of these requests uses up valuable bandwidth.  A quick test we performed searching for the term designer shoes, resulted in 11 requests and which totalled just over 100KB in the JSON responses (that doesn’t include extras such as any maps or images that may also be returned for irrelevant predicted results), a search for golf clothing gave 9 requests and 85KB.  Now these figures on their own are not massive, but if you take into account the amount of searches done on globally on google everyday (over a billion) and if each of these is resulting in say extra 50KB of internet traffic then that works out at approximately 46 Terabytes of extra traffic on the internet every day*.

So as this new feature is designed to speed up searching, could it have the opposite effect on the rest of internet browsing? It will be interesting to see whether or not all this extra traffic will slow down the rest of the web.