Tag Archives: kindle fire

Amazon Silk: fast cloud-powered browser, or a new way to mine your data?

Amazon announced its new range of Kindle devices today and the web is buzzing with debate about the impact of the new Android-based Kindle Fire tablet on Apple and others.

Amazon knows how to pile high and sell cheap, and can make money from content even if it gives away the hardware, so it is a strong contender in this space.

The real innovation announced today though was in the web browser. Amazon announced Silk, which splits the browser between your Kindle Fire and EC2 (Elastic Compute Cloud).


Amazon’s point: it can hold a massive cache of web content on EC2, as well as performing common-sense optimizations like scaling images to an appropriate size before sending them to your device.

Is this really new? Much of it sounds familiar, if you know about caching and proxies. Nevertheless, Amazon is in a strong position with its large cloud resource, and can design the web browser specifically for its cloud proxy. In addition, it knows the exact size and capability of the device. And perhaps its smart engineers have come up with better ways to cache. One feature is predictive caching – sending down the page it things you will visit next, before you actually go there.

There are some hard problems, as I have found in trying to optimize my own web site. Caching dynamic content, so that PHP script does not get executed by every browser request, is an obvious thing to do; but web pages draw content from multiple sources, including scripts that serve ad content that is meant to be targeted for the specific viewer. Optimizing that is harder.

It does also occur to me that a side-effect of Silk is that every single bit of browsing you do will go through Amazon and could potentially be mined for data about your browsing habits. Amazon, naturally, is well-placed to send you related ads from its own retail site. Amazon has not mentioned this aspect, but I am sure it has been thought about.