The Internet of Things (IoT) is one of those evolving security areas that presents, as most things do, both opportunity and risk. Security professionals will tend to view the threats that it presents through their previous experiences, even though the threat models may be different.
Solutions that have worked in the past are usually the first ones used to deal with current problems. Such solutions, though, may be the wrong answer to a problem; even though they can be a generally good and valid solution.
Take encryption, for example. If you are dealing with an IoT device, you may insist upon the messages that it sends and receives be encrypted in some manner. The encryption is assumed by most to guarantee security.
But, what if there was a way around the encryption itself? Some researchers at Princeton came up with a paper entitled "Privacy Attacks and Defenses on Encrypted IoT Traffic" that assumed the case of traffic made up of encrypted messages and still showed how such traffic could still give up important data.
They looked at a "smart home" full of IoT devices. Basically, they realized that the metadata of the traffic would allow a passive network adversary to infer private in-home user activities using just the streamed traffic rates and packet headers.
The IoT things will talk to their various servers. A DNS request is one example of this. Traffic rates (which are patterns of repetitive queries to the device's remote management servers) along with the first six digits of MAC addresses generated may also provide much information. Since the MAC addresses won't be visible on ISP-level traffic, a LAN tap will come in handy in order to grab them.
Spikes in traffic can also infer when IoT devices are active. Such spikes will have nothing to do with the actual encrypted content issued by the device, but correlate to how it is being used.
There is a footprint that will be left by IoT devices whenever they are connected, and this aspect must be considered by the security team. Though the paper uses the example of a smart home, an organization will leave the same kinds of correlatable operational traces by the use of its own IoT devices.
This is the kind of threat avenue that a simple threat model may never consider. Yet, because of its passive nature and its simplicity it may prove to be rather effective.
The researchers have an idea what to do here. They basically want to upset the correlations that could be otherwise derived by inserting dummy network requests at random intervals. Called "traffic shaping," achieving it is not usually a drain on bandwidth as might be feared. For instance, using it on combined audio and video streaming only adds 40KB/s to the overall bandwidth requirements.
This traffic shaping idea may not be the total answer to IoT privacy by any means. But it is a step in the right direction, and shows the kind of innovative and wide-ranging approaches that need to be taken to the IoT for both security and remediation.
- Big IoT Security Benefits From Service Providers Thinking Small
- Programmed to Kill: The Risk of Hacked Robots Is Real
- How Secure Are Your IoT Devices?
— Larry Loeb has written for many of the last century's major "dead tree" computer magazines, having been, among other things, a consulting editor for BYTE magazine and senior editor for the launch of WebWeek.