IIC Journal of Innovation 2nd Edition | Page 11

Toward a Safe and Secure Medical Internet of Things 3. PRIMARY ANALYSIS Our foremost objective towards laying down the foundation to secure clinical environments was to identify security risks, threats and requirements of various clinical scenarios. These are listed in the table below. Our findings have been mostly consistent with some of the existing literature on the topic [6][7] as far as external attackers are considered. However, we found that an important yet often neglected requirement is to minimize the impact of insider attacks posed by already-compromised devices that are unknowingly used in ICE settings. We discuss such an attack to instantiations of ICE that utilize secure transports such as TLS in Section 4. Attack Class Description Susceptible Components Destroy Physically destroy ICE components; e.g. cut an infusion pump tube. All Architectural Components of ICE Disturb Modify exchanged data to prevent All Architectural Components of ICE correct operation of components; e.g. man-in-the-middle or replay attacks Reprogram Modify data or code in an ICE component to prevent its correct operation; e.g. modify infusion pump software to deliver extra medication All Architectural Components of ICE Except the Communication Network Itself Denial of Service Exploit bugs or interfaces that were not designed with security in mind All Architectural Components of ICE Eavesdrop Communication Network Listen in on the deployed ICE environment to learn sensitive information. Table 1. General attack model for ICE as identified in [6] Use of an ICE controller based on DDS Security potentially addresses or mitigates Disturb, Denial of Service and Eavesdrop attacks. Further, it would mitigate the impact of insider attacks dramatically. - 10 - June 2016