digital knowledge. digital culture. digital memory.

25.9.07

webmail and low bandwidth DE


New to this blog? Why not subscribe to its feed or sign up for free email updates?


In my last blog2blog post to Robert W Martin, I asked him to explain why he wants to live his life online without needing his own equipment - something he calls digital existence.

You gave a pretty decent answer, Rob. For me, one of the most powerful motives for seeking digital existence is that it is now technically conceivable to overcome the security concerns inherent in hardware independence. This point was hammered home by your comment on your experience with electronic health records. If Alberta Health and Wellness can provide secure remote access to legally protected patient records, certainly it is possible to provide secure (or at least secure enough) remote access to my CV drafts, letters to the power company, and weiqi game records?

Digital existence has a different attraction in the South Pacific. If I had to characterize computer use in this part of the world it would be as follows:

  • There is a large cohort of older, wealthy, professionals who are heavy Internet users at home and at work; they can afford the high price of connectivity
  • There is a larger cohort of young users who have very little disposable income
  • This younger cohort are primarily unsophisticated but passionate users of low bandwidth social computing - hi5, bebo, facebook, free sms gateways to local mobile phone companies, and photo sharing sites
  • This younger cohort would love to share files and video as well, but the slow and/or expensive connections in the region make this impractical - you can almost hear a chant of I want my youtube
  • This younger cohort does not own their own computers nor do they typically have Internet connections at home - they rely on Internet caf├ęs, computer labs at educational institutions, and their workplace to get online
In short, 20-somethings in the South Pacific are living low-bandwidth digital existence right now. However, they are doing so with very little understanding if the privacy and security ramifications of their activities.

So, with that, time for some the paranoia. When we started this, you asked, "can you really trust webmail?" A great question. Let's examine this with a little differential risk analysis. If you were going to send me an email, the list of locations where your message falls under threat would be as follows:
    The traditional POP3/IMAP (i.e. Outlook Express) scenario:
    Rob's POP3/IMAP client, Rob's PC, Rob's LAN, Internet, Rob's mail server, followed by the Internet again and then into an area influenced by my email choices.
And the webmail scenario looks like this:
    The webmail (i.e. gmail, yahoo, or hotmail) scenario:
    Rob's browser, Rob's PC. Rob's LAN, Internet, Rob's webmail host, followed by the Internet again and then into an area influenced by my email choices.
Let's agree that the risks inherent in your message traversing your PC, your (possibly wireless) LAN, the Internet, and my email-sphere-of-influence are common to both scenarios and mention them no further. Let's focus on the two legs of the journey that differ:
  • POP3/IMAP client vs. browser
  • the POP3/IMAP/SMTP mail server vs. the webmail server (which includes SMTP of course)
Looking at the clients, I think a well chosen mail client is no more or less secure than a well chosen browser. Both can operate with or without SSL/TLS (if supported by the server), both can render HTML and can execute Javascript, and both are extendible with various privacy and security enhancing plug-ins. They differ in that the mail client saves all of your mail on a local drive, which is great if you are the only user or a machine but terrible if the machine is used by multiple users. I suppose you could rig your mail client to store your mail on a removable device.

The browser, on the other hand, will often write some or all of your webmail fetched mail to cache - especially on a shared computer where you do not control the settings . Even once the cache is cleared, your mail may linger until some cryptographic disk wiping takes place, unless you cache to a removable device. Also, your browser would be vulnerable to session hijacking attacks that would not impact a mail client. For machine independent secure emailing, a thumb drive mounted mail client and a thumb drive mounted browser (see xb browser) are probably equally good, but having a thumb drive feels like cheating when the point was to have no hardware of your own. If you disallow thumb drives, the browser seems to come out ahead in the digital existence balance.

Looking at the servers, both traditional POP3/IMAP/SMTP servers and webmail server's can archive some or all of your email after it has been sent or received, including messages that you have deleted. Perhaps the difference is that webmail servers are guaranteed to have a copy of all of your mail, and it will be all indexed and ready for searching by:
  • you
  • any data mining software
  • any advertising (think gmail) software
  • any unscrupulous sysadmin
  • any criminal who gains access to this juicy repository of information
  • any government agent with a warrant (Patriot Act or otherwise)
Still the use from anywhere nature of webmail is invaluable to the goal of digital existence. So the conversation naturally moves towards cryptography...

Photo by: nico.cavallotto

1 comments:

Omar Cruz said...
This comment has been removed by a blog administrator.