Why I moved my static blog to HTTPS

Fialka ciphering machine
Photo credit Brett Neilson

By the year 2015 general awareness of the differences between plain HTTP and protected HTTPS seems to become more or less general. When making online purchases, or sending private messages, or entering passwords online we are now getting this warm fuzzy feeling if there is green lock icon shining next to the page's URL. And panic attacks if there isn't one. Being HTTPS only became standard for online services dealing with customer sensitive data.

On the other hand there are pages like personal blogs, or, well, Wikipedia. Or BBC. Just good old unencrypted HTTP. You can actually access Wikipedia via https, but if you try the same trick with BBC page, it will just redirect you back to the land of plain HTTP. And why would you want encrypted connection anyway? It's not like they are selling stuff and you're submitting personal credit card number. You're just wandering around the Web, reading stuff, minding your own business.

Recently I moved my personal blog pages to be HTTPS only. Here are few reasons why.

  1. With plain HTTP there is no certain way of telling if you are actually reading my blog. You type URL in the browser (or click a link), and browser politely requests content for this address. During the journey of this request, at any point it can be read, interpreted, and responded with any content. And the browser will display whatever is returned.

  2. A classic example would be man-in-the-middle attack. Open unencrypted traffic allows anyone on your (free) wifi network to intercept it, track your requests, as well as potentially alter content of the page. With example of BBC pages, you could inject some news articles, or in general alter the content in any possible way.

  3. If you think that only evil hackers would do such atrocities, think again. There are reports of in-flight and hotel wifi networks injecting banners into webpages. Goes without saying that this would not be possible with HTTPS pages.

  4. Google started using HTTPS support on the page as a positive ranking signal last year. Meaning, HTTPS pages will be ranked higher than plain HTTP only pages.

  5. HTTPS traffic is much harder to block and filter out with corporate firewalls. You would have to force install additional certificate onto machines in that network (which is a common practice in certain places).

  6. In general, living in post-Snowden era, when TheVerge reveals new details on digital surveillance almost weekly, strong reliable encryption of everything seems like a very good idea.

  7. One interesting topic is state controlled firewalls, like Great Firewall of China. Wikipedia article on this matter contains following curious trivia:

The Tor anonymity network was and is subject to blocking by China's Great Firewall. The Tor website is blocked when accessed over HTTP but it is reachable over HTTPS so it is possible for users to download the Tor Browser Bundle.

The very same Google recently announced that not all SSL certificates will be treated equal, and SHA-1 certificates are going to be retired and indicated as unsafe already in Chrome 41. In a way, it is a danger in itself to blindly trust any SSL certificate. There are quite a few online tools to test various security aspects of your SSL - the best and most helpful being SSL Test. It will test pretty much everything, and provide you with a grade and bunch of hints how to improve it, and why it is capped (if it is).

One notable thing about SSL Test is that once you decide to buy a certificate, you can run this test against pages of the issuer, to make sure that it is indeed legit, and there are no issues like broken certificate chain.

This post is inspired by HTTPS everywhere talk on Google IO 2014, and this ATP podcast episode.

Alex Savin

Software engineer

More from Alex