Google's Chrome browser has been heavy hit by people claiming it has serious privacy issues.

After an interesting discussion with people who are, “relevant to my interests”, the topic of Google Chrome soon came up. Most predictably what was mentioned was the privacy issues raised with it. As a believer in the open-source vetting process, I heavily disagree that Google could be doing anything subversive to our personal data.

To settle this argument in a non-subjective way I decided the best course of action would be an experiment. Essentially, what I wanted to test was the difference in packets between Google Chrome on first boot-up and search term “test123″ and a comparable browser with similar conditions. This way, we test Chrome against a vetted control browser. A few conditions before we get started:

• Experimental Browser: Google Chrome Beta for Linux (Ubuntu Karmic 9.10 x64) with experience reporting disabled.
• Control browser: Midori 0.2.3 (Ubuntu Karmic 9.10 x64), a WebKit browser licensed under a similar permissive license as Chrome.
• Analysis Method: A tcpdump, sudo tcpdump -w testbrowser.pcap -s 1550 dst www.google.com, command and Wireshark analysis.
  • This tcpdump commands monitors all traffic going to the Google domain (216.239.32.0 – 216.239.63.255)
  • Wireshark was to make the dump look pretty.
• Procedure: Start the dump, load up the browser and enter a search term to Google for “test123″. Stop the packet dump after this and monitor the results.

What I found:

The results actually surprised me in a very positive way. Not only did the two browsers communicate on exactly the same ports but Chrome also used around 60 less packets in the final dump. A cruel twist of irony that a third-party browser communicates with Google over 100% more than Chrome does. Anyway, without further hesitation, here is the dump from the different conditions.