I hesitate to post commentary about this article, but feel that is important to deconstruct claims by those believing they have all the answers. Especially when they are posting in high-profile blog sites.
InfoWorld’s Security adviser Roger A. Grimes has detailed his “Perfect Plan” for making the Internet secure for every user. In his words: “All computer devices, users, and transactions must be authenticated by default.”
My husband has been working on a pet project lately that needs to have a user login system. Although he could build one himself or purchase a system, he is probably going to go with OpenID. Using OpenID simplifies the project immensely and is probably more secure than anything he or I could write. Already it is estimated that there are over 160-million OpenIDs with nearly ten-thousand sites supporting OpenID logins (http://openid.net/what/). But it does beg the question, how secure is OpenID?
OpenID is “an open, decentralized, free framework for user-centric digital identity (http://openid.net)”. Basically, a user sets up an account with one of several OpenID Providers (openid.net, aol.com, etc.). The provider keeps the username, password, email and all sorts of other account information the user wants there. When the user goes to a site that uses OpenID authentication (blogger.com, lol.com, and more), they enter their OpenID and are redirected to the Provider’s site. Here they enter their credentials and grant access to the referring website. That is the process in a nutshell, but see this video for a really great, succinct explanation.
The driving idea behind OpenID is to have only one set of credentials for all your online identities. This way you do not have to remember which username goes to which website and passwords for each. Sounds pretty good… but what happens if your OpenID is compromised. An adversary has access to ALL your online accounts. The consequences of a compromised OpenID are intense. On the other hand, people generally use the same username and password for everything anyways, which is definitely a security problem and has the same consequences of a compromised OpenID.
Benefits of OpenID are that small businesses and developers do not need to implement their own login system, users can change personal information or passwords once and have it apply everywhere, and users are less likely to do dumb things like write lists of usernames and passwords.
However, OpenIDs have some problems also. First, OpenIDs are URLs- for example, http://inkblotpassword.com/id/jessica. For an average user, a URL is difficult to remember and very unfriendly. Personally, I think users would get used to it just as they have with email addresses. There is nothing innately harder about URLs. The OpenID system is prone to phishing attacks because the user is redirected to the provider’s page which could easily be imitated. There have been problems with CSRF attacks (cross site request forgery attacks). One of the largest providers, MyOpenID.com site, had this issue, but when notified, they reacted promptly. Another issue is that the set of specifications that a provider must implement is fairly small. There are no requirements on the strength of passwords or even to have a password. From a security standpoint, OpenID just adds another layer of complexity for things to go wrong. It also puts a burden on the user to choose a provider they can trust.
With all this in mind, is OpenID a good system? Will it prove to be the downfall of the Internet as some naysayers have speculated? Or will it bring about a revolution in convienence? Should a website use OpenID as their username and password management system? Would it be an acceptable system for banks or other financial institutions?