Efficient aggregation of encrypted data in Wireless Sensor Network

1. Efficient aggregation of encrypted data in Wireless Sensor Network Author: Einar Mykletun, Gene Tsudik Presented by Yi Cheng Lin Date: March 13, 2007

2. Outline • Introduction • Conditions • Additive Homomorphic Encryption • Aggregation • Result • Conclusion

3. Introduction • Security in Wireless sensor networks (WSNs) is a popular research topic in recent years • Aggregation techniques are used to reduce the amount of data communicated within a WSN and thus conserves battery power • This paper blend inexpensive encryption techniques with simple aggregation methods to achieve very efficient aggregation of encryption data

4. Conditions(1/2) • Multi-level network tree (3-ary) • Additive homomorphic encryption M: message space C: ciphertext space M is a group under operation C is a group under operation c1 = Enck1(m1), c2 = Enck2(m2)

5. Conditions(2/2) • Computing the average and variance • The packet header is 56 bits • End-to-end aggregation in WSNs • Compare to hop-by-hop (HBH) and No-Agg

6. Additive Homomorphic Encryption

7. Additive Homomorphic Encryption • n different ciphers ci • M >= >= with t = max(mi) • The key stream k can be generated by using a stream cipher, such as RC4 t

8. Aggregation • Computing the Average • cxi = Enc(xi, ki, M),M = n*t, log(M)=log(n)+log(t) • Cx = • Sx = Dec(Cx, K, M) = Cx – K (mod M), where K = • Avg = Sx/n • Computing the Variance • yi= xi2, cyi= Enc(yi, ki’, M’),M’ = n*t2 ,log(M’)=2*log(n)+log(t) • Cy = • Vx = Dec(Cy, K’, M) = Cy – K’ (mod M),where K’ = • Variance = Vx/n – Avg2

9. Result

10. Conclusion(1/2) • This scheme is slightly less bandwidth efficient than the hop-by-hop aggregation scheme • However it provides a much stronger level of security • One limitation of this proposal is that the identities of the on-responding nodes need to be sent along with the aggregate to the sink

11. Conclusion(2/2) • Reduce n (number of nodes) or t (value)