Secure Data Aggregation in Wireless Sensor Networks: A Survey

Yingpeng Sang, Hong Shen Yasushi Inoguchi, Yasuo Tan, Naixue XiongProceedings of the Seventh International Conference on Parallel and Distributed Computing,Applications and Technologies (PDCAT'06)Presented by kevin wang

Preview

Main contributions Outline Classify by infrastructure in WSNs Classify by encryption in WSNs Proposed two general schemes

Hop by hop End to end

Conclusions

Main contributions

Past Only focus on data confidentiality or data integrity

Now Survey the work

Hop-by-hop End-to-end

Propose security frameworks respectively for Hop-by-hop End-to-end

Both on Data confidentiality and Data integrity

What is confidentiality

Confidentiality Ensuring that information is accessible only to

those authorized to access One of the cornerstones of Information security The delivering data is confidential in WSNs

For avoiding to leak secret information, the sensed data have to encrypt to keep confidentiality

M Enk(M) Dnk(M)=M

Sensor or aggregator sink node

What is integrity

Integrity Ensuring that only authorized parties are able to

modify computer system assets and transmitted information

One of the cornerstones of Information security The delivering data is sensitive in WSNs

For avoiding to modify the secret information, the sensed data have to keep integrity

Especially, in a cheaper and simple device

Outline in this paper

A survey paper for data aggregation in WSN Proposed two data aggregation scheme for

HBH and ETE respectively

Problem definition

How to satisfy the confidentiality and integrity in WSN

Classify with Infrestructure

Wireless sensor networks HWSN

Hierarchical Wireless Sensor Networks DWSN

Distributed Wireless Sensor Networks

Server

Header

Sensor NodesSacrificed NodeSensor Nodes

Server

Classify with Data aggregation

Hop-by-hop Adv: deliver package size small Disadv: key management

Pair wise key dist. DWSN Group wise key dist. HWSN

perform operators: sum, min, max, avg, count, median…

Classify with Data aggregation End-to-end

Adv: the secrets share between sink and sensor Disadv: much redundant are sent Can not perform above operators

The sensed data have been encrypted

Background-network model-HWSN

Server

Header

Sensor NodesSacrificed Node

S

A

R

F

A

Background-network model-DWSN

Sensor Nodes

Server

S

s

R

Background-security requirements Confidentiality

Eavesdropping Compromised node’s key Using the compromised node’s keys to deduce all secret

information in entire network Using the compromised key to inject unauthorized

malicious nodes in network. Integrity

Injecting arbitrary chosen malicious data into the compromised S.

Modifying, forging, or discarding messages in the compromised A and F.

Background-aggregation functions Sum Average Median Minimum Maximum Count

11( ,... ) /i nin if s s s n

11( ,... ) i nin if s s s

11 1( ,... ) , ( 1) / 2,and ,... has sortedi nin r nf s s s r n s s

1( ,... ) min{ 1... }n if s s s i n

1( ,... ) max{ 1... }n if s s s i n

1( ,... ) { 1... }n if s s s i n

Hop-by-hop encrypted data aggregation in WSN

1.Security bootstrapping 1.1Pair-wise key distribution DWSN (confidentiality)

Master key based solution [14] All nodes use one key

Pair-wise key pre-distribution solution Each node shares one key with sink

Random key pre-distribution solution [10] [7] Using key ring to find one common key

Key pre-distribution schemes with deployment knowledge [15][10] DDHV’s scheme

Other solution [5][9][16]

Hop-by-hop encrypted data aggregation in WSN

1.Security bootstrapping 1.2Group-wise key distribution HWSN

(confidentiality) Symmetric group-wise key distribution [2],1992

A symmetric key can be generate among t nodes

Asymmetric group-wise key distribution [18], 2004 ECC EC-public/private

Hop-by-hop encrypted data aggregation in WSN

2.Data integrity Some related work assume that confidentiality is protected by pre-

deployed key. [12], L. Hu and D. Evans, “Secure aggregation for wireless

networks”, In Workshop on Security and Assurance in Ad hoc Networks, Jan 2003.

[18], A. Mahimkar, T. S. Rappaport, “SecureDAV: A Secure Data Aggregation and Verification Protocol for Sensor Networks”, Proceedings of IEEE GlobalTelecommunications Conference (Globecom) 2004,Nov, 2004, Dallas, TX, USA.

[21], B. Przydatek, D. Song, and A. Perrig, “SIA: Secure Information Aggregation in Sensor Networks”,In Proc. of ACM SenSys 2003, 2003.

Secure aggregation for wireless networks, 2003 Node A, deployment, symmetric pair-wise key, KAS,

RA=reading data from node A Data transmission phase

KASi=E(KAS, i)

Parent node B and aggregated result =Aggr MAC(KAS

i,Aggr) Data validation phase

R will verifies the final aggregated results using the pair-wise keys

Lower communication cost Vulnerable

Nodes, aggregators, forwarding nodes are easy to be compromised

KASi

MAC(KASi,Aggr

)

Sum(Aggr)

SecureDAV: A Secure Data Aggregation and Verification Protocol for Sensor Networks, 2004

Using Merkle Hash Tree to improve [12] Data transmission phase

A: MAC (KASi=E(KAS, i), RA)

Parent node B and aggregated result =Aggr, generate a hash value of RA by Merkle Hash function: H(RA)

Aggregator sends MAC (Aggr, H(RA, i)) to sink node, R Data validation phase

R will verifies the final aggregated results using the pair-wise keys and queries the aggregators what hash values did they sent

The queries is to check individual readings Vulnerable

high communication cost

KASi

MAC(KASi,Aggr

)

Sum(Aggr)

SIA: Secure Information Aggregation in Sensor Networks,2003 It can engage an interactive proof with the

aggregator and check whether the aggregator result is correct.

Key point Their correct build on the related trust

KASi

MAC(KASi,Aggr

)

Sum(Aggr)

Consequence

Communication cost [21]<[18]<[12]

End-to-end data aggregation in WSN

Network-wise key distribution Master key based solutions, 2005, CEG[6], 2005, CDA[11] Public key based solution, 2006[19]

Data integrity Compared to HBH, there is no efficient scheme to protect integrity

in ETE In [23], 2004, each node sends its reading to R using ETE, The R employs truncation and trimming on the RA

’s to achieve robust aggregation result against spoofed sensor.

Proposed two frameworks for data aggregation in WSN-HBH

Framework 1: Hop-by-hop encrypted data aggregation 1.The bootstrapping phase

For controlled environment HWSN, group-wise key can be generated for all nodes within each cluster

For uncontrolled environment DWSN, Pair-wise key can be distributed among each pair of sensor node

2.The aggregator selection phase R can select aggregators to construct a transmission structure

with minimum energy cost

Proposed two frameworks for data aggregation in WSN Framework 1: Hop-by-hop encrypted data

aggregation 3.The data aggregation phase

EKai,A(xi)A:(DKai,A(xi)):sum then R

4.The data transmission phasec EKai,A(xi)+MHT(EKsi,R, (xi))

5.The data integrity verification phase R hashes all (EKsi,R) to check again

Decrypt (EKsi,R) and aggregate to check correct?

Consequence

Framework 1. Confidentiality

For HWSN group-wise key For DWSN Pair-wise key

Integrity Merkle Hash Tree

Proposed two frameworks for data aggregation in WSN-ETE

Framework 2: End-to-end encrypted data aggregation 1.The bootstrapping phase and the aggregator selection

phase For HWSN and DWSN use network-wise public key K

2.The data aggregation phase Using ECC-ElGamal to aggregate and reach homomorphic

encryption

Proposed two frameworks for data aggregation in WSN Framework 2: End-to-end encrypted data

aggregation 3.The data transmission phase

Noses will commit all (EKsi,R,(xi)) of its children by MHT to R

4.The data integrity verification phase R check the commitment hash of all (EKsi,R,(K))

Consequence

Confidentiality network-wise public key K

Integrity Merkle Hash Tree

Security analysis

Compromised some nodes, R will detect with Merkle hash tree

Compromised some aggregators, R will detect with Merkle hash tree

Compromised some nodes and aggregators, R will not detect with Merkle hash tree

HBH more efficient than ETE HBH less secure than ETE, in compromised

some nodes.

Conclusions

Survey and classify the related work into HBH and ETE data aggregation scheme

Proposed two schemes for data aggregation in HBH and ETE, respectively.

Comment

Good Know the data aggregation field Research history

More This schemes did not consider the no response nodes

problem Consider MST + dynamic routing path to reduce the end-to-

end communication cost to increase entire lifetime Past did not consider nodes will be exhausted, then have to

change path in end to end environment.