Robust Networking Architecture and Secure Communication Scheme for Heterogeneous Wireless Sensor Networks

Robust Networking Architecture andSecure Communication Scheme for

Heterogeneous Wireless Sensor Networks

McKenzie McNeal IIIPh.D. Candidate for Computer & Information Systems Engineering

Advisor: Dr. Wei Chen

College of Engineering, Technology, and Computer ScienceMarch 15th, 2012

1

Outline

Research Background and Challenges Problem Statement Research Goal and Objectives Key Related Work Conceptual and Preliminary Design Detailed Design and Implementation

Robust networking architecture Secure communication scheme

System Evaluation and Test Results Evaluation of robust networking architecture Analysis of secure communication scheme

Benchmarking Conclusion & Recommendations

2

Research Background – Wireless Sensor Networks (WSNs)

Large collection of small wireless devices with the ability to sense, process, and transmit data. Low cost solution to distributed applications• Military• Civilian

Limited resources• Power• Storage• Processing• Communication

Unreliable communication Unattended operation• Operate autonomously

Homogeneous or Heterogeneous 3

Low-end node (L-node)

Homogeneous Wireless Sensor Network

H

H H

High-end node (H-node)

Low-end node (L-node)

Heterogeneous Wireless Sensor Network (HWSN)

General security concerns for communication networks Data needs to be protected Unauthorized access Protection against various attacks

Specific security concerns for WSNs Resource constraints do not support

traditional security methods Attacks can drain network resources Uncontrollable/hostile environment

4

Research Background – Security Concerns for WSNs

Research Challenges

Network Infrastructure Reliability and availability High performance Leverage security tasks

Secured Data Communication Data confidentiality, integrity, freshness & authentication WSNs do not support traditional security methods Function in presence of node compromise

5

Key Related Work

6

Reference Network Architecture

Security Model Limitations

LIGER Flat HWSN Hybrid key management scheme (LIGER)•Unbalanced key distribution•LION-standalone key mgmt.•TIGER-KDC based key mgmt.

•Large numbers of keys stored and key exchanges•Increased node compromise with increased key storage

Kejie Lu Flat HWSN 2 key management schemes•Random key pool-based pre-distribution•Polynomial-based pre-distribution

•Large numbers of keys strored and key exchanges•No analysis for energy usage

Du-Scheme Hierarchical Region-based HWSN

Key management scheme•C-neighbor concept•ECC supports exchange of symmetric key

•Location dependent network architecture•No energy analysis for secure routing

Key Related Work (cont’d)

Summary of LimitationsNo security oriented network hierarchyRandom key pre-distribution schemes encounter the key

exchange issue Large storage of pre-loaded keys Large number of key exchanges

Localization information needed for establishing network architecture

No energy analysis for secure routingResilience against node compromise w/o tamper resistant

hardware

7

Problem Statement

Novel security methods and models are needed for HWSNs to function in the presence of an attack. Heterogeneity provides hierarchy that leverages resource efficient security tasks. This dissertation research focuses on developing a robust networking architecture and secure communication scheme with an efficient key management system and secure routing protocol.

8

Research Goal and Objectives

GoalAddress security challenges and develop a robust networking architecture and secure communication scheme for HWSNs with resource saving key management system and provide secure data communication and resilience against node compromise.Objectives

Define and develop robust hierarchical heterogeneous networking architecture Design secure communication scheme based on the defined hierarchical

HWSN Key management system Cryptographic algorithms Secure and efficient routing protocol

Test and evaluate robust networking architecture and secure communication scheme

9

Conceptual Design

Security system that integrates a robust networking architecture and secure communication scheme for HWSNs

10

Security System for

HWSNs

Secure Communication Scheme

Robust Networking Architecture

Conceptual Design

Efficiency of computation – computation of cryptographic keys and data encryption should be fast

Efficiency of communication protocol – data routing/relay should have low latency

Efficiency of energy – computation and communication tasks for security should not drain the limited power of the sensor nodes

Long Network lifetime – networking architecture can be reconfigured

11

Performance Requirements

Conceptual Design

Data confidentiality –secure channel to prevent information leakage

Data integrity – data should not be altered when transmitted from node to node

Data freshness – data should be up-to-date w/o any replay of old messages

Authentication – verify identity of source Availability – preserve energy while providing

security Self organization –robustness to overcome

node failures and node compromise 12

Security Requirements

HH

H

What is the optimal way to design robust hierarchical networking architecture to support resource efficient security for HWSNs? 13

SINK

Flat HWSN: Data transmission by flooding

HH

Send Data Back

Hierarchical HWSN: Data transmission by hierarchical architecture

Conceptual Design – General Idea

14

H

H H

H-nodeSINK

L-node

H

H H

H-nodeSINK

L-node

H

H H

Cluster

Cluster-head

Cluster member

H-nodeSINK

L-node

H

H H

Cluster

Cluster-head

Cluster member

H-nodeSINK

L-node


Data routing/relay

Self-Formatio

nReconfigurati

on

Conceptual Design – Proposed Cluster-based Hierarchical Networking Architecture

(CHNetArch)

Complete graph

15

H-node

Cluster-head

(L-node)

Cluster member(L-node)

H-node

Shared Key

Public key


Design

Secure Routing Protocol

Key Management System

Cryptographic Algorithms

Key Pre-distribution Scheme

Key Management

Protocol

Public Key Cryptograph

y

Shared Key Cryptograph

y

Conceptual Design – Proposed Secure Communication Scheme

Detailed Design and Implementation – Robust Networking Architecture

General Assumptions Communication range: H-node (D) and L-node (d) Algorithms run in rounds.

Each round consists of 1 transmission, 1 reception, and data processing

Data Structures H-node: list of L-nodes in its region, parent and children on

the backbone tree L-node: cluster head, region head

Cluster head: its cluster member list, the parent and children on the backbone tree

CHNetArch

Data routing/rela

ySelf-

FormationReconfigurati

on

Construction of CHNetArch

16

17

CHNetArchSelf-

formation

Node Move-outHead Rotation Node Move-in

Detailed Design and Implementation – Robust Networking Architecture (CHNetArch)

L-node

H

H

SINK

H

H

H

H

H

HH

H

SINK Cluster head

Cluster member Regional

headStep 1 – Algorithm for region formationRound 1H-nodes broadcast their IDs and L-

nodes receive H-nodes IDs and select H-node with strongest signal

18

Self-formation of CHNetArch

Region head

H-nodeStep 2 – Algorithm for cluster

formationA – Neighbor discoveryRound 1

L-nodes broadcast their IDs and receive IDs

B – ClusteringRounds 1 - 4

L-nodes form clusters by choosing the neighboring node with the lowest ID to be its cluster head

Step 3 – Algorithm for BT formationA – Regional backbone treesStart at region head: region head

becomes activeRounds 1 – 3(1) The active nodes find children,

then turn to inactive(2) Then the children become activeThe above process repeats until the

regional backbone tree is completeB – Connect Regional backbone

treesSink and regional heads form a tree

rooted at the Sink in the same way as regional backbone tree formation


19

Theorem 1Given a heterogeneous wireless sensor network (HWSN), its cluster-based hierarchical networking architecture (CHNetArch) can be formed in O(T) rounds, where T is the height of the backbone tree of CHNetArch.


20

CHNetArchreconfigurati

on

Node Move-outHead Rotation Node Move-in


21

Head RotationRound 1 – 2

Head request remaining energy.Cluster members send back energy amount.

Round 3 – 5Head chooses new cluster headHead informs cluster members and parent

and children on backbone tree of new head, then changes status to cluster member

Cluster members, parent and children update new head

Reconfiguration of CHNetArchNode Move-inA – Join as cluster memberRound 1 - 2

New node broadcasts a message to join at range d/4 and receives replies

Round 3 New node chooses a cluster head with strongest

signal and becomes cluster memberB – Join as cluster headRound 3 – 5

New node broadcasts a message to join at range d and receive replies

New node chooses a parent (cluster head with weakest signal)

Node Move-outA – Leaving node is cluster member

Rounds 1 – 2Cluster member sends message to cluster head and receives reply, then leaves network

B – Leaving node is cluster headRounds 1 - 7Cluster head invokes head rotation then follows steps to leave network as cluster member


22

Theorem 2The reconfiguration of CHNetArch can be done in O(k) rounds, where k is the maximum number of neighboring nodes for an L-node.


23

Data Routing/Relay


u

H

Sink

cluster member

cluster head

regional head

Data relay starts at u: u becomes active. Round 1-2(1) The active node transmits the data to its parent, and becomes inactive. (2) The parent becomes active.The above process continues until the data reaches its final destination

Theorem 3The data routing/relay of CHNetArch can be done in O(T) rounds, where T is the height of the backbone tree in CHNetArch.

24

Security System for

HWSNs



Detailed Design and Implementation – Secure Communication Scheme


Design

Secure Routing Protocol

Key Management

SystemCryptographic

Algorithms

Key Pre-distribution

Scheme

Key Management

Protocol

Public Key Cryptograph

y

Shared Key Cryptograph

y25


Public-key cryptography Elliptic Curve Cryptography

(ECC) Elliptic Curve Integrated

Encryption Scheme (ECIES) Used for public key encryption

and decryption Elliptic Curve Digital Signature

Algorithm (ECDSA) Used for authenticated

broadcasting between region head and cluster head

26


H-node

Cluster-head

(L-node)


H-node

Shared KeyPublic key


Shared-key cryptography Symmetric key generation using bivariate polynomial

x and y are IDsaij are large prime number

coefficients t is degree of the polynomial, where t

is 50

27


Security Property It requires t compromised nodes to attach the symmetric keys generated by bivariate polynomial

H-node

Cluster-head

(L-node)


H-node

Shared Key

Public key


Key management protocol Type of keys KG – pre-loaded temporary global

symmetric key K(x)pb/K(x)pr – public and private key

for node x Kuv – symmetric key shared between

node u and v, Kuv = Kvu

Broadcast message {sender.id, key(sender.id, [message])}

Unicast message {sender.id, receiver.id, key(sender.id,

receiver.id, [message])}28

Key Management System


Key pre-distribution scheme H-nodes Temporary global symmetric key ECC private/public key pair

L-nodes Temporary global symmetric key Private key of ECC pair

H-node

Cluster-head(L-

node)


H-nodeKG

Shared Key

Public key

Key distribution along with CHNetArch self-formationPurpose: (1) Guarantee network architecture formation is

secure(2) Distributed keys will also be used for secured

data routing/relayHow to distribute the keys? (3) In region formation, K(H)pb (encrypted by KG) is

broadcasted to all L-nodes.(4) After the backbone tree is formed, Each region

head H sends L-node list (encrypted by K(H)pr) in its region to the basestation/sink.

(5) The basestation sends the public key list (encrypted by K(H)pb) of the L-lodes to region head H.

29

Key Management Protocol

H

H

HH

H

SINK Cluster head

Cluster member Regional

head


CHNetArch self-reconfiguration Key used for reconfiguration:

Kuv – symmetric key shared between nodes u and v Head rotation, node move-in, and node move-out use

Kuv for any transmission from u to v Sender: {u.id, v.id, Kuv(u.id, v.id, [message])} Receiver decrypts message using Kvu and compare plaintext

(u.id, v.id) with encrypted text (u.id, v.id)

30

Key Management Protocol


31

Key used: K(H)pb/K(H)pr – public and private key of

region head K(u)pb/K(u)pr – public and private key of cluster

head or cluster member Kuv – shared key between u and v H-node to H-node

{H1.id, H2.id, K(H1)pr(H1.id, H2.id, [message])} Cluster head to H-node

{u.id, h.id, K(u)pr(u.id, h.id, [messasge])} Cluster member to cluster head

{u.id, v.id, Kuv(u.id, v.id [message])}

H2

H1

v

uuv

Secure Routing


32

1 2 3 4 5 M-1 M…

Timeslot

…i

Encrypt DecryptTransmit

1 Timeslot

TDMA Used for broadcasting during region formation Number of H-nodes known Assigned fixed timeslots

MAC Protocol

System Evaluation and Test Result– Evaluation of Robust Networking Architecture

33

…0 1 k-1 …0 1 k-1 …0 1 k-1 …0 1 k-1

Frame 1 Frame 2 Frame 3 Frame 4

Transmission in a random timeslot Receive

Encrypt DecryptTransmit

1 Timeslot

Timeslot

CSMA/CA Used for unicast Nodes transmit at random timeslot in each frame

MAC Protocol


34

Total packet size: 28 bytes Initialization vector (IV)

Destination (DST) Active message type (AM) Length of message (LEN) Source (SRC) Counter (CTR) – 216 different messages

Encrypted data Data

MACode (also known as MAC) – check integrity

DST(2 bytes)

AM(1 byte)

LEN(1 byte)

SRC(2 bytes)

CTR(2 bytes)

DATA(16 bytes)

MACode(4 bytes)

IV Encrypted Data

Data Packet Structure and Size


35

Proposed AM Types for CHNetArch FormationDescription Name Sender to Receiver

Initialize region formation INTRF Sink/basestation to all nodes

Region formation RFMSG H-node to L-nodes

Neighbor discovery NDREQ L-node to L-node

Clustering (head request) CHREQ L-node to L-node

Clustering (head replying confirmation) CHREP L-node to L-node

Clustering (head drops member request) CHDREQ L-node to L-node

Clustering (head drops member reply) CHDREP L-node to L-node

Backbone tree formation (find children) BTREQ H-node to L-node L-node to L-node

Backbone tree formation (replying to parent) BTREP L-node to H-node

Backbone tree formation (confirm from parent) BTCFM H-node to L-node


Formulas based on clustering algorithm and MAC protocols were used to evaluate the time complexity and energy consumption for CHNetArch formation and reconfiguration

36

Functions Time Energy

H-node L-node H-node L-node

Transmission THTx TLTx EHTx ELTx

Reception THRx TLRx EHRx ELRx

Listening THLx TLLx EHLx ELLx

Sleep/Idle THSx TLSx EHSx ELSx

Symmetric Encryption THSE TLSE EHSE ELSE

Symmetric Decryption THSD TLSD EHSD ELSD

Asymmetric Encryption THAE TLAE EHAE ELAE

Asymmetric Decryption THAD TLAD EHAD ELAD

Variables used for evaluation of time complexity and energy consumption


Time complexity for region formation

TRF – the time it takes to complete region formation THSE – the time it takes an H-node to perform symmetric encryption TLRx – the time it takes an L-node to receive a message TLSD – the time it takes an L-node to receive a message

Energy consumption for region formation

ERF – the total energy consumed during region formation EHSE – the energy consumed by an H-node to perform symmetric encryption EHTx – the energy consumed by an H-node to transmit a message ELRx – the energy consumed by and L-node to receive a message ELSD – the energy consumed by an L-node to perform symmetric decryption

37

Examples of formulas for CHNetArch formation


38

Communication operation(one packet of 28 bytes/108 kbps)

Energy cost (mJ) Time (ms)

Transmit 134.4 2.07Receive 150.8 2.07Listen 8885.7 131.05

Type of node Storage (KB) Encryption / packet Decryption / packetRAM ROM Time (ms) Energy (mJ) Time (ms) Energy

(mJ) MICAz 2 10 1.53 39.08 3.52 89.90

Type of node Storage (KB) ECIES (Encryption) / packet

ECIES (Decryption) / packet

RAM ROM Time (ms) Energy (mJ) Time (ms) Energy (mJ)

MICAz 1.774 20.768 3907.48 98.78 2632.66 63.18

Type of node Storage (KB) Verify (mJ)RAM ROM Time (ms) Energy (mJ)

MICAz 1.51 19.308 61800.34 58480

MACode Storage (KB) Energy (mJ)RAM ROM

CMACode 1 5.8 387.19

Time and energy consumption for communication operations on MICAz

Storage, time, and energy consumption for using AES-128 on MICAz

Storage, time, and energy consumption for using ECC on MICAz

Storage and energy consumption for using ECDSA on MICAz

Storage and energy consumption for using MACode: CMACode

Sensor node modeling


500 x 500 meter sensor field 20 H-nodes 1000 – 3000 L-nodes (increments of 500) H-nodes communication range: D = 250 meters L-nodes communication range: d = 60 meters

39

Simulation Environment


40

Number of clusters

1000 1500 2000 2500 30000

100

200

300

400

500

600

700

800

Number of L-Nodes

Num

ber

of c

lust

ers

Average size of a cluster

1000 1500 2000 2500 30000

0.5

1

1.5

2

2.5

3

3.5

4

4.5

Number of L-Nodes

Ave

rage

size

of a

clu

ster

Number and size of clusters in CHNetArch


41

Execution time

1000 1500 2000 2500 30000

10000

20000

30000

40000

50000

60000

Number of L-Nodes

Exe

cutio

n T

ime

(sec

onds

)

1000 1500 2000 2500 30000

50000

100000

150000

200000

250000

Number of L-Nodes

Ene

rgy

(Jou

les)

Energy consumptionTime and Energy consumption for CHNetArch formation


42

1000 1500 2000 2500 30000

10000

20000

30000

40000

50000

60000

Region DiscoveryClusteingBackbone Tree Formation

Number of L-Nodes

Exe

cutio

n T

ime

(sec

onds

)

1000 1500 2000 2500 30000

50000

100000

150000

200000

250000

Region DiscoveryClusteringBackbone Tree Formation

Number of L-Nodes

Ene

rgy

(Jou

les)

Execution time Energy consumption

Time and Energy consumption for each phase of CHNetArch self-formation


43

Percentage of e/E, where e is the energy used for CHNetArch formation, and E is the total energy amount for two AA batteries in each L-node

1000 1500 2000 2500 30000.012%

0.013%

0.014%

0.015%

Number of L-Nodes

Perc

enta

ge o

f e/E


44

1000 1500 2000 2500 30000

200

400

600

800

1000

1200

1400

Head rotationNode move-in as cluster memberNode move-in as cluster headNode move-out as cluster memberNode move-out as cluster head

Number of L-Nodes

Ene

rgy

(Jou

les)

1000 1500 2000 2500 30000

5000

10000

15000

20000

25000

Head rotationNode move-in: clus-ter memberNode move-in: clus-ter headNode move-out: clus-ter memberNode move-out: clus-ter head

Number of L-Nodes

Exe

cutio

n T

ime

(sec

onds

)System Evaluation and Test Result

– Evaluation of Robust Networking Architecture


Time and Energy consumption for each phase of CHNetArch reconfiguration

45

1000 1500 2000 2500 30007863.5

7864

7864.5

7865

7865.5

7866

7866.5

7867

7867.5

7868

Number of L-Nodes

Exe

cutio

n T

ime

(sec

onds

)

1000 1500 2000 2500 3000760

770

780

790

800

810

820

830

Number of L-Nodes

Ene

rgy

(Jou

les)



Time and Energy consumption for data routing/relay

46

System Evaluation and Test Result– Analysis of Secure Communication Scheme

Evaluation of Key Management SystemThe following variables help define the number of keys stored in CHNetArch

Nh – number of L-nodes in a region where h is region head Kh – number of neighboring H-nodes of an H-node h Nch – number of cluster members in a cluster where ch is cluster head Kch – number of neighbors on backbone tree for cluster head ch Nc – number of clusters in CHNetArch, which is same as number of cluster

heads Let Ah be number of keys stored by a regional head:

Let Bch be the number of keys stored by a cluster head:

Let Ccm be the number of keys stored by a cluster member:

47


Evaluation of Key Management System

Let Ah be number of keys stored by a regional head: Let Bch be the number of keys stored by a cluster head: Let Ccm be the number of keys stored by a cluster member: Let Kall be the total number of keys stored in CHNetArch:

Variable DefinitionNh number of L-nodes in a region where h is region head

Kh number of neighboring H-nodes of an H-node h

Nch number of cluster members in a cluster where ch is cluster head

Kch number of neighbors on backbone tree for cluster head ch

Nc number of clusters in CHNetArch, which is same as number of cluster heads

The table of variables help define the number of keys stored in CHNetArch

48

Number of stored keys

1000 1500 2000 2500 30000

2000

4000

6000

8000

10000

12000

14000

16000

Number of L-Nodes

Num

ber

of k

eys



49

Memory needed to store security algorithms and keys on a cluster head and cluster member

1000 1500 2000 2500 300056300

56400

56500

56600

56700

56800

56900

57000

cluster membercluster head

Number of L-Nodes

Mem

ory

(KB

)

Cluster member two 160-bit keys for ECC one 128-bit shared key

Cluster heads Two 160-bit keys for ECC One 128-bit shared key with each

cluster member One 128-bit shared key with

backbone neighbors For symmetric polynomial

q = 296

(t + 1)log2 q = 0.612 KB 44% of memory use for security



50

Security AnalysisProvides data confidentiality

Public key and shared key cryptographyProvides data freshness

Counter in IV ensures at least 216 different messagesProvides data integrity

MACode computer over data packet can be verified by receiverProvides data authentication

Sender and receiver IDs are sent in plain text and encrypted text Compare for verification


Benchmarking

51

Comparison of pre-loaded keys

1000 1500 2000 2500 30000

10000

20000

30000

40000

50000

60000

70000

Proposed key preload-ing: approach 1Proposed key preload-ing: approach 2Du-scheme

Number of L-Nodes

Num

ber

of p

re-lo

aded

key

s

Proposed System vs Du-scheme Networking architecture

Self-formation No location information

Key management Pre-loads less keys Stores less keys

Secure communication Uses temporary global

symmetric key

Conclusion

Robust networking architecture (CHNetArch) Performs self-formation without location information Nodes communicate according to hierarchical network architecture Backbone tree provides high networking performance Network architecture is reconfigurable

Secure communication scheme Resource saving key management system Combination of public and shared key cryptography for secure

network formation Secure routing protocol governed by network hierarchy and key

management system Provide resilience against node compromise

52

Security System for HWSNs

Recommendations

Storage of security algorithms for ECC can be reduced by adjusting switches used for calculation on sensor nodes

Symmetric bivariate polynomial can be designed for a larger value of t

Provides increased resilience against node compromise Use more than one symmetric bivariate polynomial

i.e., one for each region Further research can be conducted to find resource efficient

methods to provide security for HWSNs

53

Security System for HWSNs

Systems Engineering Management PlanSEMP

54

Research Activity 2009 2010 2011 2012SP SU FA SP SU FA SP SU FA SP SU F

AConceptual DesignNeed AnalysisFeasibility StudyPreliminary DesignSystem RequirementsSystem DecompositionTechnical Performance MeasuresProposed SolutionDetailed DesignModeling of Network ArchitectureNetworking Architecture Formation and Reconfiguration AlgorithmsKey Management Protocol and Cryptography

ImplementationNetwork ArchitectureKey management ProtocolMAC ProtocolsTesting and EvaluationTest and Evaluation ModelNetwork PerformanceSecurity AnalysisBenchmarkingReport WritingConference and Journal PublicationsDissertation Final Report

Publications

1. McNeal III, M., Chen, W., Shetty, S., and Aungst, S., “Joint Design of Cluster-Based Hierarchical Network Architecture and Key Management System for Heterogeneous Wireless Sensor Networks”, IJCES International Journal of Computer Engineering Science, Volume 1 Issue 3, pages 49-66. December 2011.

2. McNeal III, M., Chen, W., Shetty, S., and Aungst, S., “Security-Oriented Robust Networking Architecture and Key Management for Heterogeneous Wireless Sensor Networks”, 10th International Conference on Wireless Networks, 2011.

3. Liang Hong, McKenzie McNeal III, Wei Chen, “Secure cooperative MIMO communications under active compromised nodes”, 9th IEEE International Conference on Pervasive Computing and Communications Workshops, 2011.

4. Wei Chen, McKenzie McNeal III, Liang Hong, “Cross-Layered Design of Security Scheme for Cooperative MIMO Sensor Networks”, 2010 IEEE International Conference on Wireless Information Technology and Systems, 2010.

5. Long, K.J., S.E. Haupt, G.S. Young, L.M. Rodriguez, and M. McNeal, “Source Characterization using a Genetic Algorithm and SCIPUFF”, Seventh Conference on Artificial Intelligence and its Applications to the Environmental Sciences at AMS Annual Meeting, Phoenix, AZ, 2009.

55

Acknowledgements

Committee Members: Dr. Wei Chen Dr. Sachin Shetty Dr. Mohammad Bodruzzaman Dr. Ali Sekmen Dr. Liang Hong Dr. Stanley Aungst

College of Engineering, Technology, & Computer Science Dean Hargrove Dr. Malkani

PSU Research Team Dr. Sue Haupt Kerrie Long Andrew Annuzio Luna Rodriguez

Defense Threat Reduction Agency (DTRA) DTRA01-03-D-0010-0016 56

Questions/Comments

57

问题 /评论 ?Wenti/Pingrun

58

Homework and assignment

1. How can the cluster-based networking architecture in a sensor network leverage the efficiency of security system?

2. How to realize data confidentiality in a flat sensor networks? Consider a key system.

3. Discuss the tradeoff between public cryptograph and private cryptograph on power, storage, processing and communication, respectively.

4. Give the definition of a compromised node in sensor networks. Does the security system here detects compromised nodes?

Robust Networking Architecture and Secure Communication Scheme for Heterogeneous Wireless Sensor Networks

Documents

key exchangesno analysis

heterogeneous3lowend

communication networksdata

secure routingresilience

outlineresearch background

computer sciencemarch

dissertation research

low cost solution