Asymmetric Homomorphic Encryption Transformation for Securing Distributed Data Storage in Wireless Sensor Networks

In asynchronous wireless sensor networks it is not possible to transmit the monitored data to an authorized recipient in real-time. Thus, the data needs to be stored within the network in distributed fashion. However, due to the resource constraints of the sensor nodes the amount of data to be stored and transmitted must be minimized. One popular technique for this is in-network data aggregation. But because wireless sensor networks are employed in public environments, the data stored in the network has to be protected against the attacks. As solution for asynchronous wireless sensor networks tiny persistent encrypted data storage (TinyPEDS) has been proposed. TinyPEDS provides in-network aggregation of encrypted data and stores the aggregated data in a replicated and distributed manner. Since the data being aggregated is encrypted, the encryption scheme needs to support an additive homomorphic operation on ciphertexts. This thesis studies the implementation and integration of such an encryption scheme, namely the elliptic curve ElGamal encryption (EC-ElGamal) in TinyPEDS. In order to keep the resource consumption at a minimum, several algorithms for operating at finite field and elliptic curve level are analyzed and compared according to their timing values and code sizes. Several trade-offs and their effects on memory usage, code size and code efficiency are discussed. Moreover, a new approach for reducing multiprecision addition over GF (p) is proposed. In order to enhance the performance and decrease the unnecessary overhead in resource consumption, several techniques are employed which ensure an appropriate programming style.