Monday, July 18

BrowserID a Threat to Individual Freedom?

The folks at Mozilla recently introduced BrowserID. You can compare it to OpenID, but there are some key differences. The basic idea - a single set of authentication credentials across multiple sites and simplified logon to each as facilitated by the browser. Ian Yip took an interesting look at BrowserID from the an Identity Management industry perspective and how it relates to what we call identity federation. For more details on how it works, check Lloyd Hilaiel's post.

But that's not what I wanted to write about. I'm more interested in SC Magazine's article headlined Mozilla BrowserID "seriously flawed" and Roger Clarke's Reaction to Mozilla's BrowserID Proposal, which was the subject of the SC Mag article. My first point is simply: go read it. It gives you a lot to think about. My second point, though, is a little more complex.

Clarke makes some interesting and compelling arguments about Internet privacy and individual freedom. I can't say that his logic is incorrect or that his points are invalid, because they're not. But his anger (characterized by phrases like "seriously flawed 'identity management' schemes" and "its design is seriously threatening to individual freedoms") may be a bit misplaced.

I agree with Richard that BrowserID is not THE solution to solve the Internet's authentication and privacy problem. But that's not the challenge that Mozilla has sought to solve. Not every site that requires a logon is a major privacy risk. I have probably 50 or more web site accounts to manage and I welcome solutions to my credential management problem. I'm a security guy but I will gladly introduce some level of risk to make life easier when browsing a large number of those sites. We all do to a degree. e.g.) It's less risky to go to a library anonymously to look something up in a book but the Internet at home is just so much more convenient that we risk being eavesdropped or introducing malware to our systems every time we use it.

It reminds me of the old argument that two-factor authentication is useless because it's susceptible to MITM attacks. BrowserID won't be a silver bullet for all authentication scenarios and maybe not even for ANY scenarios that require high security or strong assertions about the user, but it could still be a useful way for end-users who want to simplify the logon process. Claiming that BrowserID is seriously flawed because it doesn't address issues outside of its own scope just seems wrong and even somewhat irresponsible. The IT industry's version of media sensationalism maybe?

I don't mean to pick on SC Mag - the title got me to read Richard's article, which is the purpose of a strong title, but I'm pulling for one of these solutions (OpenID, CardSpace, BrowserID) to make it into the mainstream so that my life will be a little easier. And creating hysteria and FUD around them doesn't help with user adoption.

No comments: