Protocols for Secure Communication in Wireless Sensor Networks

More documents

Recommendations

Info

126 Chapter 4. Key Establishment Based on Blom’s scheme, an optimized version for wireless sensor networks has been proposed in [61]. Two modifications have been made: first, not only one key space but multiple spaces are being used (i.e. multiple Blom schemes are run in parallel), and second, the matrix G is not explicitly given but instead generated from a seed (which saves storage space). Out of the ω key spaces generated, the KDC chooses τ of them randomly for each node. Therefore, this approach is a combination of random key-predistribution with the deterministic Blom scheme. Two nodes can establish a shared key if they have at least one key space in common. This is the complementary probability of the event that two nodes share no key space. It is given by 1 − ωω−τ τ τ ω2 τ = 1 − ((ω − τ)!)2 (ω − 2τ)!ω! The resilience of the scheme can be expressed as follows. Assuming x nodes have been compromised, the probability that a pairwise key of two uncompro- mised nodes is broken is pκ = x ∑ j=k+1 x τ j 1 − j ω τ x− j ω Key agreement based on hash chains Hash chains were introduced by Lamport [107] for the verification of one-time passwords. First, a hash chain is generated from a seed. The seed is secretly given to the user, while the last element of the chain is given to the host where the user wants to log on. In order to log on, the user sends the value of the chain that hashes to the value that is stored in the host. If both values match, access is granted and the host stores the value just sent. This procedure continues until the seed has been used for logging on. This method has the advantage that passwords are safe even if they are overheard by an enemeny, since each password is used only once. The one-way property of h ensures that verification of passwords is possible, while it is infeasible to derive the next password to log on from the current one. Key agreement based on multiple hash chains has been described in [112]. The authors describe a more general variant first, where purely random values instead of hash-computed ones are distributed to the participants. This provides unconditional security, but has the disadvantage that large data sets have to be distributed. They then develop the version based on hash-computed values, which is more efficient but provides only computational security that is based on the infeasibility to inverse one-way functions.
4.7. Summary 127 The security of the scheme is discussed, as in our work, in terms of the probability with which an attacker can eavesdrop on a pairwise key. While they derive upper bounds for this probability, we provide an exact combinatorial derivation as well as an approximation that is well-suited for practical purposes. Combination of random key pre-distribution and hash chains The combination of the hash-based scheme with random key pre-distribution schemes has been independently described by Ramkumar and Memon [148], who call this scheme “hashed random preloaded subsets” (HARPS). Using a similar methodical approach, they arrive at basically the same conclusions as us. 4.7 Summary The tight resource contraints on nodes in wireles sensor networks, both in computational and storage terms, make traditional key management mechanisms based on public-key cryptography largely impractical. Nevertheless, security goals like resilience against eavesdropping, impersonation, and the creation of fake identities should be achieved. On the other hand, wireless sensor networks are constrained in several regards that make it possible to rely on certain assumptions when designing appropriate key management mechanisms. It can be assumed that all nodes in the network originate from the same administrative domain, the set of nodes is known before deployment and relatively static throughout the network’s lifetime, the number of secure associations per node is small compared to the overall size of the network, and the contribution of a single node to the overall functionality of the network is also relatively small. The key agreement schemes discussed in this chapter are designed to be applicable in wireless sensor networks as their resource consumption is very low. They require a moderate memory size, which is independent of the network size, and only computationally cheap operations, such as hash functions. The fact that nodes are within one administrative domain means that for authentication, nodes have to simply prove that they belong to this domain, which makes it possible to rely on a probabilistic authentication scheme. Pre-distributing key material before deployment is easily possible since it is known in advance which nodes are going to be deployed together. Under the assumed node capture attack model, it has to be assumed that a certain fraction of nodes is under control of the attacker. As a consequence, it has to be assumed that some of the reported data by sensor nodes is potentially manipulated, a fact that has to be dealt with on the application layer.
Page 1:
Diss. ETH Nr. 18174 Protocols for S
Page 4 and 5:
ii often ad hoc and transient natur
Page 6 and 7:
iv Dieses Ziel kann durch kryptogra
Page 9 and 10:
Contents 1 Introduction 1 1.1 Backg
Page 11 and 12:
Contents ix 3.3.1 Basic Assumptions
Page 13 and 14:
Contents xi 6.3 Performance Evaluat
Page 15 and 16:
Chapter 1 Introduction 1.1 Backgrou
Page 17 and 18:
1.3. Contributions 3 on the other h
Page 19 and 20:
1.4. Thesis Outline 5 • Mario Str
Page 21 and 22:
Chapter 2 Wireless Sensor Networks
Page 23 and 24:
2.1. Applications of Wireless Senso
Page 25 and 26:
2.1. Applications of Wireless Senso
Page 27 and 28:
2.2. Sensor Node Characteristics 13
Page 29 and 30:
Page 31 and 32:
Page 33 and 34:
Page 35 and 36:
2.3. Related Network Types 21 zero,
Page 37 and 38:
2.3. Related Network Types 23 • P
Page 39 and 40:
2.3. Related Network Types 25 PAN,
Page 41 and 42:
2.4. Related Device Types 27 in con
Page 43 and 44:
2.4. Related Device Types 29 ports
Page 45 and 46:
2.5. Sensor Network Models 31 both
Page 47 and 48:
2.5. Sensor Network Models 33 are a
Page 49 and 50:
2.6. Simulation of Sensor Networks
Page 51 and 52:
Page 53 and 54:
Page 55 and 56:
2.7. Security Requirements 41 terac
Page 57 and 58:
2.7. Security Requirements 43 netwo
Page 59 and 60:
2.7. Security Requirements 45 acces
Page 61 and 62:
2.7. Security Requirements 47 2.7.6
Page 63 and 64:
2.7. Security Requirements 49 of te
Page 65 and 66:
2.8. Cryptography for Sensor Networ
Page 67 and 68:
Page 69 and 70:
Page 71 and 72:
2.9. Existing Approaches to Wireles
Page 73 and 74:
Page 75 and 76:
Page 77 and 78:
Page 79 and 80:
Chapter 3 A Security Model for Wire
Page 81 and 82:
3.1. Attack Paths 67 field (cf. [17
Page 83 and 84:
3.1. Attack Paths 69 modules [63].
Page 85 and 86:
3.1. Attack Paths 71 closed environ
Page 87 and 88:
3.2. Attack Objectives 73 • Timel
Page 89 and 90: 3.2. Attack Objectives 75 Critical
Page 91 and 92: 3.2. Attack Objectives 77 Actors A
Page 93 and 94: 3.3. Adversary Characteristics 79 a
Page 95 and 96: 3.3. Adversary Characteristics 81 c
Page 97 and 98: 3.3. Adversary Characteristics 83 p
Page 99 and 100: 3.4. The Cost of End-to-End Securit
Page 101 and 102: 3.4. The Cost of End-to-End Securit
Page 103 and 104: 3.5. Approximating End-to-End Secur
Page 109 and 110: 3.7. Summary 95 Location-constraine
Page 111 and 112: Chapter 4 Key Establishment Shared
Page 113 and 114: 4.2. Random Key Pre-Distribution 99
Page 125 and 126: 4.3. Key Agreement Based on Hash Ch
Page 131 and 132: 4.4. Multiple Hash Chains for Key A
Page 133 and 134: 4.4. Multiple Hash Chains for Key A
Page 135 and 136: 4.5. Strengthening Random Key Pre-D
Page 137 and 138: 4.6. Related Work 123 pc = 0.33 Du-
Page 139: 4.6. Related Work 125 Probability o
Page 143 and 144: Chapter 5 Multipath Communication T
Page 145 and 146: 5.1. Principles of Multipath Commun
Page 147 and 148: 5.2. Routing on Spanning Trees 133
Page 155 and 156: 5.3. Properties of Tree Paths 141 b
Page 157 and 158: 5.3. Properties of Tree Paths 143 a
Page 159 and 160: 5.3. Properties of Tree Paths 145 n
Page 161 and 162: 5.3. Properties of Tree Paths 147 F
Page 163 and 164: 5.3. Properties of Tree Paths 149 D
Page 165 and 166: 5.4. Security Evaluation 151 compro
Page 167 and 168: 5.5. Related Work 153 width usage a
Page 169 and 170: 5.5. Related Work 155 separation. F
Page 171 and 172: Chapter 6 Integrity-Preserving Comm
Page 173 and 174: 6.1. Authentication and Integrity P
Page 175 and 176: 6.2. Basic Interleaved Authenticati
Page 189 and 190: 6.3. Performance Evaluation 175 A B
Page 191 and 192:
6.3. Performance Evaluation 177 We
Page 193 and 194:
6.4. Security Evaluation 179 Bandwi
Page 195 and 196:
6.4. Security Evaluation 181 a sens
Page 197 and 198:
6.4. Security Evaluation 183 Number
Page 199 and 200:
6.4. Security Evaluation 185 We ass
Page 201 and 202:
6.4. Security Evaluation 187 Paths
Page 203 and 204:
6.4. Security Evaluation 189 ity of
Page 205 and 206:
6.4. Security Evaluation 191 of the
Page 207 and 208:
6.4. Security Evaluation 193 The se
Page 209 and 210:
6.5. Extended Interleaved Authentic
Page 211 and 212:
Page 213 and 214:
Page 215 and 216:
Page 217 and 218:
Page 219 and 220:
Page 221 and 222:
6.6. Comparing Interleaved and Mult
Page 223 and 224:
6.7. Applications 209 Psi 1 0.8 0.6
Page 225 and 226:
6.7. Applications 211 codes are att
Page 227 and 228:
6.7. Applications 213 domain. The r
Page 229 and 230:
6.9. Summary 215 simulations and by
Page 231 and 232:
Chapter 7 Conclusion This work prop
Page 233 and 234:
7.1. Secure Communication in Wirele
Page 235 and 236:
7.2. Future Work 221 7.1.4 Shortcut
Page 237 and 238:
Bibliography [1] Specification of t
Page 239 and 240:
Bibliography 225 1st European Works
Page 241 and 242:
Bibliography 227 [41] Rohan Chitrad
Page 243 and 244:
Bibliography 229 [64] L. Eschenauer
Page 245 and 246:
Bibliography 231 [85] Yahoo! Inc. D
Page 247 and 248:
Bibliography 233 [105] Kristof Van
Page 249 and 250:
Bibliography 235 [126] Anne Marie M
Page 251 and 252:
Bibliography 237 [150] Sylvia Ratna
Page 253 and 254:
Bibliography 239 [172] Thanos Stath
Page 255 and 256:
Bibliography 241 [192] Eric W. Weis
Page 257:
Curriculum Vitae Harald Vogt Person
show all

Protocols for Secure Communication in Wireless Sensor Networks

Create successful ePaper yourself

Delete template?

Save as template?