Each of these hardware features poses some inherent risk. The utility of these features is predicated entirely on standardization -- which means each of these technologies is readily available to potential malicious actors.
In most cases, it is not practical to reinvent data storage or transport technologies as alternatives to USB, WiFi, BT, or Ethernet. Typical strategies for reducing these risks include device pairing, password protection, data encryption, and data format. Though the scope of this article does not allow an in-depth exploration of all possible mitigation measure, we’ll review several of these briefly.
Device pairing is a method (familiar from BT) that requires specific instances of a device (i.e. a particular sensor, a particular WiFi hub, a particular PC, etc.) and limits access to the device only to explicitly paired devices. This strategy can be layered on top of communication between any pair of devices where the protocol already provides a handshake that shares device ID, or where it is possible to add some custom software to share and record device IDs and authentication.
It is also possible to implement pairing via a back-end service, accessible from the host device, that stores approved pairings in a single database. One advantage of a global database (similar to IMEI numbers for cellphones) is that it allows early detection of counterfeit devices (duplicate IDs). Of course, one disadvantage is that it requires a convenient connection between the host device and the back-end service, something that is not always possible or practical, especially in countries with underdeveloped infrastructure or technologically reluctant or low-income patients without easy Internet access or reliable phone service.
Passwords are a valuable tool for preventing unauthorized access, and can be implemented in a wide range of schemes that balance between usability, convenience, and security.
Data encryption might be considered passé as the increase in computing power and the increased sophistication of factorization techniques has made public key encryption methods subject to some attacks. DI Management offers a good description of some of the mathematics and weaknesses behind public key encryption and RSA, in particular. Additionally, PKI methods require additional computational power to encrypt and decrypt data that may not be practical in all circumstances.
Data formats is likely to be considered by many the most naïve method in this list. However, it is worth mentioning at least partly because it is naïve, relatively easy to implement, and can provide some deterrent for the most casual malicious actors. This can be as simple as defining custom messages to communicate via WiFi, Ethernet, ZigBee, or using a binary file format rather than a more convenient text-based file format. It could also be as elaborate as a custom file system to protect data, a custom USB class, or a new set of op-codes and an interpreter to protect company IP included in the software executable image.
This overview of four well-known techniques is hardly an exhaustive treatment of these techniques, and there are others (signing, certificates, custom bearers, biometrics, redundant authentication, physical keys) that are also valuable tools that can be used appropriately in some circumstances.
The enhancements from new technology to patient safety, treatment efficacy, and the user experience for both patients and healthcare professionals are significant. Innovative companies will find ways to incorporate these technologies to increase the value proposition of their product offerings. To be competitive, device manufacturers will need to find ways to utilize these technologies effectively and address the associated security issues. The good news is that security issues can be addressed by a combination of mitigations and product requirements.
Alan Walsh is director of northeast software engineering for Logic PD.