Miyaji and Omote, in this paper, explore the security of encrypted self-healing features for three different wireless sensor network (WSN) schemes designed to be deployed in hostile environments. In such an environment, one imagines the reverse engineering of captured nodes, the introduction of compromised or Trojan nodes, hacked communication links, and node replacement following battery failure (no joke). Healing goes beyond replacing hardware to the restoration of security with cryptographic keys.
Some of these networks rely on servers and a fixed home base, so they are not homogeneous. Eager attackers are described as regularly attacking sensors and links, possibly without disrupting network operations. They may learn the network’s topology (which is assumed to be regular, unchanging, and probably not very large) and the names of nodes. To me, a network of named sensors connected by a fixed communications architecture will be very brittle, compared to one of anonymous sensors connected by a dynamic architecture.
The authors’ investigation is focused on (1) RPoK, a resilient polynomial-based random key pre-distribution scheme (RKP); (2) S-RKP, a simple (that is, limited program) RKP; and (3) POLISH, a proactive, self-healing, cooperative link scheme without a server. Sensors and links are described as being green (fully functional), yellow, or red (compromised), as well as healing (the process of increasing green nodes). Mathematical analysis and simulation are used to compare the availability and the efficiency of healing processes. This is interesting.
They write: “The RoK scheme was the first resilient multiphase WSN with self-healing.” RoK is credited to Castelluccia and Spognardi [1]. Except for one 1993 paper, the references mention only work published after 2002. The authors do not seem to be aware of theoretical work in this area during the decades following 1970, in particular, a very large solar-powered, randomly deployed, homogeneous, anonymous, asynchronous, self-healing sensor network loosely described in works by Kotin and Stark [2,3], later depicted in the Hollywood movie Tears of the Sun [4], and finally in section 12 of Stark [5]. The only thing about this WSN not published is the very simple and possibly unbreakable security trick used when new sensors must be added to the network. None of the encryption techniques mentioned in the paper at hand are used.
This paper will be of interest and of value to engineers working with the networks described by the authors.