September 2010

Android applications are supposed to get permission from the user before they gain access to personal information, such as location. But what happens once the permission is given?

This study from Network World looked at 30 apps to see where and when personal information was used, and found some worrisome results.

A recent test of prototype security code for Android phones found that 15 of 30 free Android Market applications sent users’ private information to remote advertising servers, without the users being aware of what was being sent or to whom. In some cases, the user’s location data was sent as often as every 30 seconds.

Can the Internet be encrypted by default?

With the current debates about lawful intercept and increasing numbers of man-in-the-middle attacks, maybe the Internet should finally be made secure by default.

Encryption is currently used sparingly, mostly when connecting to e-commerce and financial services over the web. Here the https protocol is used and traffic between the user’s web browser and the server is protected from eavesdropping using SSL. The problems with this scheme are legendary, mostly associated with requiring users to notice when encryption is on and off, and knowing how to interpret certificate information and error messages.

But could encryption be turned on all the time, automatically?

Google has recently made https the default for Gmail, demonstrated that encryption can be scaled to millions of users. What about scaling it to the entire Internet?

Tcpcrypt is an extension to the TCP protocol designed to make encryption the default. It is backwardly compatible with traditional TCP, and it would protect old applications that don’t have encryption. And it works faster than the SSL we rely on today.

You can read more about tcpcrypt in a recent technical paper, on a tcpcrypt community web site, and on Wikipedia.

A good, brief article on economics and security failures by Ross Anderson contains some great quotes…

The discipline of security economics teaches us that large systems often fail because incentives are poorly aligned; if someone guards a system while someone else bears the cost of failure, then failure is likely.

…

As one of my students put it, “All the party invitations in Cambridge come through Facebook. If you don’t use Facebook you don’t get to any parties, so you’ll never meet any girls, you won’t have any kids and your genes will die out.”

Google is introducing two-factor authentication to its Google Apps products. This means that in order to access the enterprise services (mail, documents, etc.) the Google user will have to know their password and also supply a one-time verification code. That code will be sent to your cell phone, or generated by a special application on the smart phone.

The approach is not novel, and does not provide 100% security, but it is notable because of Google’s size and influence. Having such a large player adopt stronger authentication can only help to speed the adoption by other organizations, and that is a good thing.

By doing this now, and previously making https the default in gmail, Google is demonstrating that better security can be done on a large scale, with general users.

There are two research positions open at CMU in the area of privacy decision making. One is at the Post-Doc level and the other is for Ph.D. students. The principle investigator is Alessandro Acquisti.

The project aims at investigating the role of soft paternalistic approaches in assisting users who face privacy-sensitive trade-offs. Such privacy “nudges” will be incorporated into policy proposals as well as tools and technologies to be developed by other members of the project.