EFFICIENT GENERALIZED SIGNCRYPTION BASED ON ECC

International Journal on Cryptography and Information Security (IJCIS), Vol. 5, No. 2, June 2015
DOI:10.5121/ijcis.2015.5201 1
EFFICIENT GENERALIZED SIGNCRYPTION BASED
ON ECC
An Braeken1
and Pawani Porambage2
1
Industrial Engineering, Vrije Universiteit Brussel, Brussel, Belgium
2
Department of Communication Engineering, University of Oulu, Finland
ABSTRACT
In order to obtain promising wireless communication systems, it is necessary to maintain secure message
transmissions in unicast, multicast, or broadcast scenarios. The cryptographic primitives can be
incorporated when a node needs to submit messages to different receivers as encrypted and/or
authenticated one-way messages. For achieving the given objective, signcryption schemes are efficient
primitives.
In this paper, we first show some serious weaknesses of a previous proposed pairing-free signcryption
scheme. Then, we correct the scheme and extend it to a signcryption scheme with multiple users instead of
the proposed tripartite topology. The obtained signcryption scheme is further extended to a generalized
signcryption scheme. All schemes consist of elliptic curve operations and satisfy also forward secrecy and
public verification. To illustrate the usefulness of the generalized signcryption scheme, we show its
application in the context of key management for wireless sensor networks.
KEYWORDS
Signcryption, Generalized signcryption, Elliptic Curve Discrete Logarithm Problem (ECDLP), Key
management
1. INTRODUCTION
In order to obtain energy efficient message transmissions, broadcasting and multicasting
communication scenarios are widely performed in wireless network settings. Due to the open air
message deliveries and unreliable connection establishments, the wireless communication links
are always considered to be insecure. Therefore, the key security characteristics that should be
incorporated with all types of communication scenarios in wireless networks are identified as
confidentiality, authentication, and integrity. The sender needs to guarantee that the message can
only be read by authorized receivers (confidentiality) whereas, the receivers need to ensure about
the correctness of the origin (authentication) and content of the actual message (integrity). These
security requirements are efficiently achieved by a signcryption scheme.
A generalized signcryption (GSC) scheme provides either confidentiality or authentication or a
combination of both. Moreover, since a GSC advocates all these security features accompanied
by the possibility of sending multiple messages to different receivers in the same communication
phase, a GSC scheme can be deployed in constrained wireless devices also.

2
The application scenarios of signcryption schemes have gone beyond the theoretical derivations
to the real-time application level. In [1], an application for multi-receiver identity-based schemes
has been proposed in the context of Mobile Ad hoc networks (Manets) to deal with malicious
nodes. Suppose that every node has a surveillance mechanism. Then, if one node detects that its
neighbor node has malicious behavior, the node sends a warning message and broadcasts this
message to the whole Manet using the multi-receiver scheme. Once a node reaches a given
threshold of warnings, the node is excluded from the network. Other applications are related to
the key management in wireless sensor networks (WSNs) [2,3]. Signcryption is used here to share
the key between the sink node and the network nodes.
In this paper, we propose the most efficient signcryption and a GSC scheme from literature, that
use elliptic curve cryptography (ECC) primitives and avoid the computationally heavy bilinear
operations, satisfying besides the classical security properties also forward secrecy and public
verification. The signcryption scheme is an extension to multiple receivers of the tripartite
signcryption scheme from [4], where we first correct a serious security problem. This scheme is
then further extended to a GSC.
The remainder of the paper is organized as follows. Section 2 provides an overview about the
related work. Section 3 explains the key definitions and assumptions used. In Section 4, we
propose the new signcryption scheme and the extended GSC scheme. Sections 5 and 6 discuss the
security and the performance aspects of the schemes respectively. We show how to apply the
GSC scheme for key management of WSNs in Section 7. Finally, Section 8 concludes the paper.
2. RELATED WORK
The concept of public key signcryption was introduced by Zheng [5] in 1997. The significance of
a signcryption scheme is that it simultaneously generates both encryption and signature in a single
phase. Consequently, cryptographic properties such as confidentiality, integrity, authentication
and non-repudiation are more efficiently obtained compared to the traditional approaches, which
first encrypt and then sign the message. Since 1997, many variants of signcryption schemes with
different cryptographic properties have been proposed. More particularly, the identity-based
signcryption scheme was firstly introduced by Malone in 2002 [6]. In [7], Duan et. al. proposed
the inclusion of multi-receiving property to the identity-based signcryption schemes. Since then,
several other identity-based multi-receiving schemes were suggested in many previous literature
[8].
Most of the proposed signcryption systems in literature make use of the computationally heavy
operations of bilinear pairing. The first ECC based signcryption scheme was suggested in [9],
which saved about 58% computational cost and about 40% communication cost compared to the
signature-then-encryption approach. In [10], a recent overview is given of all the different
signature schemes based on ECC that were proposed in literature. In particular, they were
compared with respect to the security attributes of confidentiality, integrity, authentication,
unforgeability, non-repudiation, forward secrecy, and public verification. Papers
[2,3,11,12,13,14,15] were suggested to satisfy all the required security features of a signcryption
scheme.
After thorough evaluation of these papers, we concluded that papers [2,12,15] have exactly the
same structure. Papers [3] and [11] claim to satisfy non-repudiation and public verifiability, but

3
clearly do not since the private key of the receiver is required to perform the steps of the
unsigncryption procedure in the protocol. Reference [15] consists of a completely different
approach but also requires much more ECC operations. Finally, in [13] a serious security
weakness exists because the sender sends the parameter s to the receiver, where s=twa. Here wa
represents the private key of the sender and t is a parameter that can be computed by the receiver
in the protocol. As a consequence, the receiver is able to derive the private key of the sender.
Another signcryption scheme from 2013 satisfying all the possible required properties, not
mentioned in the review [10], can be found in [4]. However, there the number of receivers is
limited to two, which is not pragmatic in real-time wireless systems. Moreover, a serious security
problem exists in the key generation of the scheme.
To summarize, our proposed signcryption scheme is an extended and corrected version of the
scheme from [4]. We show that this scheme is more efficient than the other existing full security-
featured signcryption schemes from [2,12,15] based on basic elliptic curve (EC) operations. In
addition, we extend the signcryption scheme to a generalized one by adding one extra EC
multiplication at sender side.
3. SYSTEM MODEL AND ASSUMPTIONS
First, we discuss the security problem of the tripartite signcryption scheme of [4]. Then, we
correct the problem and extend the scheme to a signcryption scheme with multiple receivers
instead of only 2. Next, we extend this signcryption scheme to a GSC scheme with a limited
additional cost.
This section explains the key definitions and the assumptions used for deriving the proposed
scheme. We consider the sender S and a number of n receivers U={U1,...,Un}, where n≥2. The
identities of the sender and the receivers are respectively denoted by IDS and IDj with 1≤j≤n. The
ECC primitives are taken from the standardized curve parameters as given in [16,17], where G is
the base point generator of the curve of prime order p, defined in the finite field Fq, with q prime.
The sender's private and public keys are respectively denoted by dS and QS. Each entity in the
scheme has its own private-public key pair, whereas in particular, the j-th receiver's key pair is
(dj, Qj) for 1≤j≤n. The generation of these key pairs is performed by a trusted Public Key
Generator (PKG) and is explained in the next section. The node identities and the corresponding
public keys are publicly known.
Operation H is a one-way cryptographic hash function (e.g., SHA2 or SHA3) that results in a
number of Fq. The encryption algorithm can be considered as the well known AES or the
lightweight cryptographic algorithm PRESENT. It is assumed that the given cryptographic
operations and the EC parameters are identically implemented and performed at each
participating entity. The original message sent by the sender is represented as M. The ciphertext
C, is obtained by encrypting the message M using the key K, C=EK(M). The concatenation of two
messages M1 and M2 is denoted by M1|M2.
For the definition of the GSC scheme, we need to introduce two assumptions, related to the hash
function on the zero message and the encryption/decryption of a message under the zero-key. In
particular, H(0)=0 and E0(M)=M, D0(C)=C.

4
4. PROPOSED SCHEME
First, we discuss the security problem of the tripartite signcryption scheme of [4]. Then, we
correct the problem and extend the scheme to a signcryption scheme with multiple receivers
instead of only 2. Next, we extend this signcryption scheme to a GSC scheme with a limited
additional cost.
4.1. Security problem of previous scheme
The main weakness of the scheme in [4] takes place in the key generation. The private keys for
the participants are defined by the PKG, using the identity of the user and the private key of the
PKG dk. Thus, the private key of user a is derived as da=dkH(IDa). Consequently, any user is able
to derive the private key of the PKG. This also means that any user is able to establish own key
pairs in the name of the PKG. Another serious problem is that every time a key pair of a user
needs to be revoked and reset, it implies a change in either the identity of the user or the private
key of the PKG and thus also all other users' key pairs.
A last issue is that the system is called an identity-based scheme. Following the definition of
identity-based cryptography [18], the public key should consist of the identity of the user. Here,
the identity of the user is only involved in the private key. From a security point of view, an
identity-based cryptosystem guarantees the link between the public key and a given identity. In
[4], this link is created by including identity-related information any time the public key of a user
occurs. However, the question still remains how the valid pair, consisting of identity and public
key, reaches the user. An attacker can easily mislead a user by for instance placing his public key
with another identity IDo, allowing him to read a message intended to that particular identity IDo.
An attacker could also generate its own key pair, since the particular form of the private key is not
exploited in the scheme. We see two possibilities in solving this issue.
• The data containing the pairs (identity and public key of each user) is publicly published,
but a third party independently checks all changes on the content.
• The most secure way would be to transmit the paired information of identity and public
key by applying the signature scenario of the GSC scheme that we develop in this paper.
It needs to be mentioned that both approaches are less efficient than the key management in
identity-based crypto. However, identity based crypto is typically achieved by means of bilinear
operations, which we explicitly want to avoid due to their computationally heavy requirements.
Since transmission costs are fundamentally smaller (order of 1000) than the computation costs of
pairing operations, we will show that even with an extra communication phase of the required
public key information with the signature scenario of the GSC scheme, we are still able to obtain
a considerable more efficient system.
4.2. Proposed Signcryption Scheme
A signcryption scheme consists of 3 different phases. The first phase is the key generation of the
private and public keys of the participants. The second part handles the signcryption, executed by

5
the sender. In the third phase, the receivers perform the operations individually, which is referred
as the unsigncryption process. We describe these different processes more into detail.
4.2.1. Key Generation
We now correctly apply the idea for the integration of the identity in the private key as in [4]. Let
us also consider a PKG with (dk,Qk) the master private and public key pair. We also include a
parameter r, which is a random value in Fq. It allows the revocation of the user's credentials and
the issue of new credentials without changing the private master key or the identity of the user.
The private key and public key of S and the users Uj with 1≤j≤n is now defined as
dS = H(dkH(IDS|r) G) dj = H(dkH(IDj|r) G)
QS = dSQk Qj = djQk
The private keys are securely delivered to the users of the system in off-line mode. The master
public key Qk is stored in all nodes. The tuple (IDi, Qi, H(IDi|Qi)) is made publicly available for
all users Ui in the system. The last element in this tuple is to ensure the link between the identity
and the public key.
4.2.2. Signcryption
This process consists of 6 steps. Denote by M=(M1|...|Mn) the message to be encrypted for the
corresponding users of the system U=(U1,...,Un) where each Mj is intended for user Uj with 1≤j≤n.
We want to note that the multi-receiver property, exactly one message to all receivers, can also be
easily included in the algorithm by integrating the key derivation of [19] after step 3 of the
signcryption and step 5 of the unsigncryption procedure.
In order to start the signcryption process, the identities and the corresponding public keys of the
receivers (Q1,...,Qn) are queried by the sender. Let us here assume that these values are securely
stored in a public environment (cf. first approach). The following steps are then performed by the
sender.
1. The sender chooses a random parameter r1 and computes kS = r1Qk.
2. For all users Uj, with 1≤j≤n, the parameter kj is computed as kj=H(r1H(IDj)Qj).
3. The ciphertext is computed, Cj=Ekj(Mj). Denote by C=(C1|...|Cn).
4. Then, the hash h is calculated as h=H(kS|C|IDS|U).
5. Finally, we compute s=r1-hdS.
6. The message to transmit, consists now of s|h|C|IDS|U.
4.2.3. Unsigncryption
Any individual receiver Uj, with 1≤j≤n can now execute the following process.
A preliminary step before the unsigncryption can start, is to request or query the corresponding
public key QS of IDS and to check its integrity by means of the hash H(IDS|QS). Then, all required
info is available to check the signature and decrypt the message of the transmitted data
s|h|C|IDS|U.
1. The intermediate point P=sQk+hQS is first computed.
2. Now, the value kS=P.
3. In the next step, we can compute h'=H(kS|C|IDS|U).

6
4. If h'=h, the message origin and validity is verified and we can continue the process.
Otherwise, we stop.
5. Finally, the value kj=H(djH(IDj)P) can be computed.
6. Thus, the decryption leads to Mj=Dkj(Cj).
The different steps in this signcryption protocol are summarized in Table 1.
Table 1. Proposed signcryption scheme.
4.3. Generalized Signcryption Scheme
There are 3 main scenarios in a GSC scheme.
• Signcryption scenario: sender and receivers are determined and the message is encrypted
and provided with a signature.
• Signature scenario: Only the sender is determined. The message is only provided with a
signature and no encryption is performed.
• Encryption scenario: Only the receivers are determined. The message is encrypted and
signed by an anonymous sender.
We now show how the scheme from the previous paragraph with some small adaptations can be
transformed to a GSC scheme. The main trick we use is to add an extra random value r2 for which
the corresponding EC point R2=r2Qk is calculated. This point R2 is also included in the transmitted
message. As a consequence, R2 can be used as a kind of temporary public key and r2 as the
corresponding temporary private key of the sender. Thanks to the integration of this pair (r2, R2),
the encryption scenario of the GSC scheme, where the sender is unknown, can be defined.
The key generation is the same as described above. The input parameters of the GSC scheme are
determined by (IDS, dS, QS, U1, d1, Q1,...,Un, dn, Qn). We now describe the different modes.
4.3.1. Signcryption Scenario
The input of the system in signcryption scenario consists of (IDS, dS, QS, U1, d1, Q1,...,Un, dn,Qn).
The protocol is very similar as the signcryption protocol from the previous paragraph. The main
differences are
• The introduction of R2 (step 2 in signcryption mode) as replacement of QS (step 1 in
unsigncryption mode).

7
• Replacement of dS by r2 in the computation of the signature (step 6 in signcryption
mode).
• The additional transmission of R2 in the message and the corresponding inclusion into the
input of the hash to ensure integrity of the transmitted message.
The different steps are summarized in Table 2.
Table 2. GSC Signcryption Scenario.
Signcryption Unsigncryption
1. Choose r1 , kS = r1Qk 1. P =sQk+hR2
2. Choose r2 , R2 = r2Qk 2. kS =P
3. For all j, kj =H(r1H(IDj)Qj) 3. h' = H(kS|C|IDS|R2|U)
4. For all j, Cj =Ekj(Mj) C =(C1|...|Cn) 4. h' =h?
5. h =H(kS|C|IDS|R2|U) 5. kj =H(djH(IDj)P
6. s =r1-hr2 6. Mj =Dkj(Cj)
7. Send s|h|R2|C|IDS|U
4.3.2. Signature Scenario
The input of the system in the signature scenario consists of (IDS, dS, QS, 0,...,0), since the
receivers are unknown. Here we use the assumptions that E0(M)=M, D0(C)=C and H(0)=0.
Consequently, steps 3 and 4 from the signcryption mode and steps 5 and 6 from the
unsigncryption mode, are drastically simplified in the signature scenario.
The different steps are summarized in Table 3.
Table 3. GSC Signature Scenario.
3. For all j, kj =H(0)=0 3. h' = H(kS|C|IDS|R2|0)
4. For all j, Cj =E0(Mj) =Mj C =(M1|...|Mn) 4. h' =h?
5. h =H(kS|C|IDS|R2|0) 5. kj =0
6. s =r1-hr2 6. Mj =D0(Cj)
7. Send s|h|R2|C|IDS|0
4.3.3. Encryption Scenario
The input of the system in the encryption scenario consists of (0,0,0, U1, d1, Q1,...,Un, dn,Qn),
since the sender is unknown. Due to the introduction of the pair (r2, R2) that takes over the role of
the sender's key pair (dS, QS), the encryption scenario consists of the same operations as the
signcryption scenario. The different steps are summarized in Table 4.

8
Table 4. GSC Encryption Scenario.
3. For all j, kj =H(r1H(IDj)Qj) 3. h' = H(kS|C|0|R2|U)
4. For all j, Cj =Ekj(Mj) C =(C1|...|Cn) 4. h' =h?
5. h =H(kS|C|0|R2|U) 5. kj =H(djH(IDj)P
6. s =r1-hr2 6. Mj =Dkj(Cj)
7. Send s|h|R2|C|0|U
To conclude, with an extra cost of one point multiplication (cf. step 2 at sender side in the
scheme) and the transmission of one extra ECC point, a GSC scheme is obtained from the
signcryption scheme. Moreover, it should be noticed that thanks to the introduction of an extra
random value r2, the private key of the sender is better protected. In [20], it is considered as a
weakness to make the private key dependent of only one random value. This is also the case in
the first proposed signcryption scheme, where s=r1-hdS and s,h are part of the transmitted
message.
5. SECURITY ANALYSIS
In the previous section, we proposed 2 schemes, a signature scheme and a GSC scheme as
extension of this signature scheme. The extension of the signature scheme to the GSC scheme
consists of the inclusion of an extra tuple (r2,R2=r2Qk), that takes over the role of a temporary
private and public key pair. The parameter R2 is on the one hand included in the transmitted
message. On the other hand, at the place where an operation was involved with the private key of
the sender (step 6 of signcryption procedure), the parameter r2 is used. Consequently, due to the
Elliptic Curve Discrete Logarithm Problem (ECDLP), the security of the GSC scheme is
completely determined by the security of the signcryption scheme. Therefore, we can here restrict
our discussion to the security of our first proposed signcryption scheme. Let us start with showing
the correctness of the protocol. Next, we shortly discuss the most important security properties of
the scheme.
5.1. Correctness of the protocol
There are 2 checks to be made. One check is related with the verification of the hash and the other
with the derivation of the EC point. Both are based on the same principle.
5.1.1. Verification of the Hash
We need to check if the determination of kS by sender and receiver corresponds to each other. Let
us start from the side of the receiver and show that we obtain the same calculation as the
kS = P = sQk + hQS
= (r1-hdS)Qk +hdSQk = r1Qk

9
5.1.2. Derivation of Point on EC
Again sender and receiver have a different calculation for the computation of the shared secret
key kj. Starting from the receiver leads to the following result:
kj =H(djH(IDj)P) = H(djH(IDj) (sQk+hQS))
= H(H(IDj)((r1-hdS)djQk + hdjdSQk ))
= H(H(IDj)r1Qj)
5.2. Other Security Properties
The security of the whole scheme is based on the ECDLP. It is exploited two times, first to derive
a common shared EC point and second for ensuring authentication and integrity of the message.
Let us now shortly describe the different security properties of the protocol. Again, for the same
reason as explained above, we focus on the base signcryption scheme.
5.2.1. Public Verifiability and Non-repudiation
Thanks to the particular construction of the private and public keys of the entities, a public
verifiable scheme is created. This means that any third party without knowledge of the private
key of sender or receiver(s) can validate the signcrypted message. Since anyone with knowledge
of the master public key Qk in the scheme can execute the computations as shown in the
derivation of kS above. Then, anyone can also check the value of the hash h. As a consequence,
also non-repudiation, allowing any third party to validate the origin of the message, is obtained in
the scheme.
5.2.2. Unforgeability
The scheme is also unforgeable, which means that only the intended sender can produce a valid
message-signature pair. No other entity can send a message s|h|C|IDS|U, since the values s, h and
C are either directly or indirectly linked with the given identity IDS or the private key dS of the
sender S. First of all, with respect to the public verifiable part of the protocol, being the check of
the parameter h, it is not possible due to the ECDLP for deriving the random value r1 in the
computation of a correct value for s. Secondly, also a valid value for the secret key kj is
impossible to construct without knowledge of the private key of the sender due to the ECDLP.
The random value r1 is generated in the beginning of the protocol. This value is integrated in all
the steps of the protocol where the identity or private key of the sender is involved. The first step
is in the computation of the EC point shared with the receivers. The second step is related to the
derivation of the parameter kS, which leads to the first validity check at the side of the receiver. As
a consequence, no replay attacks can be performed in order to forge a signature.
5.2.3. Message Integrity
It is clear that the message integrity is obtained thanks to the hash h, transmitted in the message.
This hash value includes information of the ciphertext C, the values ks, IDS and U. The value kS
depends on the random value r1 and cannot be replayed.

10
5.2.4. Authentication
In the signcryption scheme the authentication of the identities of sender and receivers are
obtained thanks to 2 measurements. First, the key pairs, generated by the PKG, are based on the
identities of sender and receiver. This follows from the fact that the private keys, incorporating
the identities, of the involved entities are used. Also at the moment the public key is used from an
entity, the corresponding identity is included in the operation, which creates a dependency
between the public key and the identity as is the case in identity-based cryptography. We assumed
that alternations to the list (IDi, Qi, H(IDi| Qi)) are independently verified by a third party.
5.2.5. Forward Secrecy
Even if the private key of the sender is revealed, an attacker is not able to derive the secret key kj
for the decryption of messages, sent in the past. This key kj is constructed using a random value
r1, which is deleted in the sender's memory after sending the message. The value r1 cannot be
derived from kS, due to the ECDLP.
6. EFFICIENCY
It is fair to compare our scheme with other systems in literature that achieve the same (i.e. all
required) security attributes and also restrict to basic EC operations. In any case, these systems
are much more efficient than the ones that include the computationally heavy pairing operation.
As mentioned in [21], for binary fields, pairing operations behave almost 5 times worse than EC
point multiplications operations in timing and energy performance. For prime fields, the
difference increases to a factor of 16. Consequently, one pairing operation has a huge impact in a
signcryption protocol, taken into account that our signcryption scheme with one receiver only
requires two point multiplications from sender's side.
As follows from the review paper [10] and our analysis afterwards as explained in Section 2, only
the schemes from [2,12,15] can be used for comparison. The schemes from these three papers
turn out to all have exactly the same structure.
Moreover, applying the same trick of including an extra random value r2 and a corresponding EC
point R2, makes it possible to extend their signcryption scheme also to a GSC scheme.
Table 5 compares the most computationally heavy operations between our signcryption scheme
and the ones of [2,12,15]. These operations include the EC point multiplications (PM) and point
additions (PA). We also compare the number of operations for the corresponding GSC schemes,
taking into account that the scheme of [2,12,15] is extended the same way as proposed in this
paper.
Table 5. Comparison of EC operations with n the number of receivers.
Sign PM Sign PA Unsign PM Unsign PA
[2,12,15] n+2 0 3 1
Our Signcryption Scheme n+1 0 3 1
GSC for [2,12,15] n+3 0 3 1
Our GSC n+2 0 3 1

11
As can be concluded from Table 5, our signcryption scheme requires one PM less at sender side
than the scheme of [2,12,15]. In addition, an inverse field operation is required in the schemes
[2,12,15] to compute the signature parameter s, while we only need a subtraction. The same
difference is consequently translated into the corresponding GSC schemes, where only 1 extra
PM operation is required at the sender's side. Consequently, for 1 receiver, our signcryption
scheme outperforms the other existing ones with 33% and our GSC scheme with 25%.
When we want to apply the second proposed approach, being the application of the signature
scenario of the GSC scheme, for delivering the pairs identity and public keys of the receivers to
the sender, we have to add 2 extra PM and 1 PA. As shown in [21], one pairing operation requires
around five times more energy consumption that one PM. Consequently, our scheme is able to
offer the same security objectives than an identity-based signcryption scheme and will always be
more efficient than these identity-based proposals.
7. APPLICATION IN KEY MANAGEMENT
We now shortly discuss the applicability of our GSC scheme for the key management of WSNs.
Let us assume that there is a PKG that generates the private and public keys following the method
explained in Section 4 and distributes the private keys in off-line mode to the different nodes of
the network.
There is one powerful base node (BN) that decides the cluster formation, based on the energy
level and the location. The identity and public key of the BN is also stored at each node. The BN
contains a list of sensor node identities and corresponding public keys. The following steps are
then performed in the network to manage the key distribution among the nodes.
1. The BN broadcasts an info update request to all nodes in the network. This message only
needs a valid signature from the BN.
2. Each individual node sends its info to the BN, using the signcryption scenario.
3. Based on that information, the BN calculates the candidate cluster heads. For each
cluster, the list, containing the members of each cluster and their corresponding cluster
head, is send to all nodes in that cluster. Note that in order to completely determine an
entity in the network, the identity should be accompanied with the corresponding public
key. This message is only signed.
4. The BN signcrypts a secret shared key for each of the cluster heads.
5. The cluster heads signcrypt a secret group key for the whole cluster. Note that this
corresponds with the multi-receiver signcryption scheme. The nodes first verify if the
identity of the sender corresponds with one of the nodes contained in the previous
message.
6. Suppose that every node contains a surveillance mechanism, as considered in [1]. The
node can than detect malicious behavior of its neighbor nodes and send this information
in encrypted mode to the BN.
Table 6 summarizes the number of computationally heavy operations, PM and PA, corresponding
to each step, executed by the BN, cluster nodes, and the remaining sensor nodes. Let m denote the
number of different clusters and t the number of warning messages send from the nodes of the

12
network to the BN. The total number of operations in the table corresponds with the total
executed from step 1 till 5, i.e. the key management of one single phase.
Table 6. Number of PM and PA in key management protocol using our GSC.
Step BN Cluster node Sensor node
1 2PM 2PM+1PA 2PM+1PA.
2 3PM+1PA 3PM 3PM
3 m(2PM) 2PM+1PA 2PM+1PA
4 m(3PM) 3PM+1PA 0
5 0 3PM 3PM+1PA
6 3PM+1PA t(3PM) t(3PM)
Total (5+5m)PM+PA 13PM + PA 10PM+3PA.
8. CONCLUSIONS
A very efficient signcryption scheme that satisfies all the required security features, based on the
basic EC operations, is presented in this paper. A simple trick is proposed to extend the
signcryption scheme to a generalized version with a very small additional cost of one extra PM at
sender side. Both schemes are shown to be more efficient than the others in literature. As
application of the GSC scheme, we illustrate the key management in WSNs using this protocol.
ACKNOWLEDGEMENTS
This work is supported by the TETRA grant of the Flanders agency for Innovation by Science
and Technology, and the Short Term Scientific Mission performed under COST Action IC1303.
REFERENCES
[1] Wu, L., An ID-Based Multi-Receiver Signcryption Scheme In MANET, Journal of Theoretical and
Applied Information Technology, Vol. 46, No. 1, 2012
[2 ]Esam A. A. A.H., Doaa E.S, Hazem H. A., A New Forward Secure Elliptic Curve Signcryption Key
Management (FS-ECSKM) Scheme for Heterogeneous Wireless Sensor Networks,International
Journal of Computer Science and Technology, Vol. 2, No 2, pp 19-23, 2011
[3] Bala S.,Sharma G. and Verma A.K., An Improved Forward Secure Elliptic Curve Signcryption Key
Management Scheme for Wireless SensorNetworks, Lecture Notes in Electrical Engineering
(Springer Link), Vol. 215, 2013
[4] Elkamchouchi H.M., Elkheir E.F.E. and Abouelseoud Y., A pairing-free identity based tripartite
signcryption scheme, International Journal on Cryptography and Information Security (IJCIS), Vol. 3,
No. 4, 2013
[5] Zheng Y., Digital Signcryption or How to Achieve Cost(Signature and Encryption) Cost
(Signature)+Cost(Encryption), LNCS Advances in Cryptology-CRYPTO97, Springer-Verlag Press,
1997, pp. 165-179
[6] Malone-Lee J, Identity based signcryption, Cryptology ePrintArchive, 2002
http://eprint.iacr.org/2002/098.pdf
[7] Duan S., and Zhenfu C., Efficient and provably secure multi-receiver identity-based
signcryption, Information Security and Privacy, pp. 195-206, Springer Berlin Heidelberg, 2006
[8] List of papers around signcryption (1007-2007), http://www.signcryption.org/publications/

13
[9] Zheng Y., Imai H., How to construct efficient signcryption schemes on elliptic curves,Information
Processing Letters, Vol. 68, No.5, pp. 227 - 233, 1998
[10] Singh A.K., A Review of Elliptic Curve based Signcryption Schemes, International Journal of
Computer Applications, Vol. 102, No. 6, 2014
[11] Amounas F., Sadki H., El Kinani E.H., An Efficient Signcryption Scheme based on The Elliptic
Curve Discrete Logarithm Problem, International Journal of Information and Network Security,
Vol. 2 No. 3, pp. 253-259, 2013
[12] Mohamed E., Elkamchouchi H., Elliptic Curve Signcryption with Encrypted Message Authentication
and Forward Secrecy, International Journal of Computer Science and Network Security, Vol.9, No.1,
pp 395-398, 2009
[13] Toorani M., Shirazi A.A.B., An elliptic curve-based signcryption scheme with forward secrecy,
Journal of Applied Sciences, Vol. 9, No.6, pp. 1025 -1035, 2010
[14] Ahirwal R., Jain A., Jain Y. K., Signcryption Scheme that Utilizes Elliptic Curve for both Encryption
and Signature Generation, International Journal of Computer Applications, Vol. 62, No. 9, pp. 41-48,
2013
[15] Das S., Samal B., An Elliptic Curve based Signcryption Protocol using Java, International Journal of
Computer Applications, Vol. 66, No. 4, pp. 44-49, 2013
[16] SEC 2: Recommended Elliptic Curve Domain Parameters, Certicom Research, Standards for Efficient
Cryptography Version 1.0, September 2000. [Online]. Available: http://www.secg.org/collateral/sec2
final.pdf
[17] Recommended Elliptic Curves for Federal Government Use, National Institute of Standards and
Technology, August 1999. [Online]. Available:
http://csrc.nist.gov/groups/ST/toolkit/documents/dss/NISTReCur.pdf
[18] Shamir A., Identity-Based Cryptosystems and Signature Schemes, Lecture Notes in Computer
Science, Vol. 196, pp 47-53, 1985
[19] Qin. H., Dai Y., Wang Z., Identity-based multi-receiver threshold signcryption scheme, Security and
Communciation Networks, Vol. 4, No. 11, pp 1331-1337, 2011
[20] Mohsen Toorani and Ali Asghar Beheshti Shirazi, Cryptanalysis of an efficient signcryption scheme
with forward secrecy based on elliptic curve, In Proceedings of International Conference on
Computer and Electrical Engineering (ICCEE’08), pp. 428-432.
[21] Piotr Szczechowiak, Leonardo B. Oliveira , Michael Scott, Martin Collier, and Ricardo Dahab,
NanoECC: Testing the Limits of Elliptic Curve Cryptography in Sensor Networks, European
conference on Wireless Sensor Networks (EWSN)
Authors
An Braeken obtained her MSc Degree in Mathematics from the University of Gent in 2002.
In 2006, she received her PhD in engineering sciences from the KULeuven at the research
group COSIC (Computer Security and IndustrialCryptography). In 2007, she became
professor at Erasmushogeschool Brussel (currently since 2013, VUB) in the Industrial
Sciences Department. Prior to joining the Erasmushogeschool Brussel, she worked for almost 2 yearsat a
management consulting company BCG. Her current interests include cryptography, security protocols for
sensor networks, secure and private localization techniques, and FPGA implementations.
Pawani Porambage is currently a PhD student in Department of Communications
Engineering at University of Oulu, Finland. She obtained her Bachelor Degree in Electronics
and Telecommunication Engineering in 2010 from University of Moratuwa, Sri Lanka and
her Master’s Degree in UbiquitousNetworking and Computer Networking in 2012 from
University of NiceSophia-Anipolis, France. Her main research interests include lightweight
security protocols, security and privacy on Internet of Things, and WirelessSensor Networks.

EFFICIENT GENERALIZED SIGNCRYPTION BASED ON ECC

More Related Content

Similar to EFFICIENT GENERALIZED SIGNCRYPTION BASED ON ECC (20)

Recently uploaded (20)

EFFICIENT GENERALIZED SIGNCRYPTION BASED ON ECC