Essay: Secure Group Message Transfer Stegosystem

Abstract: Grid environment is a virtual organization with varied resources from different administrative domains; it raises the requirement of a secure and reliable protocol for secure communication among various users and servers. The protocol should guarantee that an attacker or an unidentified resource will not breach or forward the information. For secure communication among members of a grid group, an authenticated message transferring system should be implemented. The key objective of this system is to provide a secure transferring path between a sender and its authenticated group members. In recent times, many researchers have proposed various steganographic techniques for secure message communications. This paper proposes a new secure message broadcasting system to hide the messages in such a way that an attacker cannot sense the existence of messages. In the proposed system, we use steganography and image encryption to hide group keys and secret messages using group keys in images for secure message broadcasting. The proposed system can withstand against conspiracy attack, message modification attack and various other security attacks. Thus, the proposed system is secure and reliable for message broadcasting.
Keywords: secure group communication, secure message broadcasting, grid environment, steganography, secure inter group communication, group key.
1. Introduction
Security is one of an important issue and essential requirement in grid environment. Grid allows users from multiple domains to work in groups by sharing information with each other. Secure group communication is applicable in various applications such as interactive simulations, multiparty military actions, government discussions on critical issues and real time information services. To secure communication among members of a group working on collaborative tasks, grid environment requires implementation of additional security mechanisms. Secure group communication in grid needs to guarantee confidentiality and validity of the message to confirm the receiver that message is forwarded by the authorized user.
The main objective of secure group message transferring system is to create a secure environment between the sender and the authorized receivers for sharing some information in a secure way. In the secure group message transferring protocol, the message communication among the sender and the receivers forming the group must be confidential. Thus, only the authorized group members can extract the message, and the unauthorized members cannot access any important information. This brings the need of secure communication protocol responsible for generating secure group key and providing authenticated secure message transferring between group members. To provide these security functions, secure communication protocol needs to generate a group key/session key for the group members based on their secret information.
Our Secure group message transferring stegosystem (SGMS) protocol is based on steganography and image encryption technique to mask secure messages in such a way that an attacker could not know that any message is being communicated in the group. This paper extends our previous work (Bhatia et al., 2013) and proposes a new secure group message transferring stegosystem that can guard the group communication in grid environment against various security attacks, such as conspiracy attack, message modification attack. As a result, the proposed stegosystem not only has advantages of the secret message transferring system, but also is more protected and realistic in comparison to already proposed message transferring system. The remainder of this paper is organized as follows: Section 2 discusses group communication protocols already proposed by various researchers. Section 3 presents brief outline of our already proposed secure group communication protocol. The newly proposed stegosystem is presented in Section 4, while Sections 5 discusses its security features. Section 6 presents the simulation and experimental results and performance, respectively. Finally, Section 7 concludes the paper.
2. Related work
Grid Security Infrastructure (GSI), a part of open grid services architecture (OGSA) provides confidentiality and security based on public key encryption, X.509 certificates and the secure socket layer(OpenSSL) for authentication and secure communication over the Internet(Foster et al., 1998).Researchers have proposed different protocols for secure group communication between grid entities. Researchers used either centralized or distributed group key management protocols for secure group communication. Most of the researchers have used encryption schemes/algorithms to provide secure group communication among various group members. Dual Level Key Management protocol (DLKM) that uses access Control Polynomial (ACP) and one-way functions to provide flexibility, security and hierarchical access control was proposed by Zoua (Zoua et al., 2007). Researchers used encryptions to update the group key for forward and backward secrecy. Li(Li et al., 2007) proposed a scalable service scheme using digital signatures and used Huffman binary tree to provide security and integrity. In this approach, Huffman binary tree is used to distribute and manage keys in VO and complete binary tree is used to manage keys in administrative domain. Park, (Park et al., 2010) have proposed an ID-based key distribution scheme that uses cryptographic algorithms to offer security and offer scalability. Li (Li et al., 2008) have proposed an authenticated encryption mechanism for group communication with basic characteristics of group communication in grid in terms of the basic theory of threshold signature. Several researchers have proposed various systems (Liao, 2007; Huang, 2010; Liaw, 1999) based upon different cryptographic techniques. Most of these systems encrypt the messages and send it to the members of its group. Liaw (Liaw, 1999) proposed a secure broadcasting cryptosystem based on the RSA and symmetric encryption algorithms, which allows addition of new users into the active groups. Tseng and Jan (Tseng and Jan, 2001) pointed out that Liaw’s broadcasting system is not secure against conspiracy attack; an intruder can break its security and can obtain the master secret key. To overcome this problem, Tseng and Jan proposed a modified broadcasting cryptosystem. Masque and Peinado (Masque and Peinado, 2006) pointed out some problems of incorrect arithmetic in Tseng and Jan’s broadcasting cryptosystem and presented a redefined Liaw’s broadcasting cryptosystem. Zhu and Wu (Zhu and Wu, 2008) showed that Masque and Peinado broadcasting cryptosystem is still insecure, in that system an unauthorized user can obtain the secret session key of other group but did not provide any solution to that redefined Liaw’s broadcasting cryptosystem.
3. Background: Proposed Protocol for secure group communication
This section briefly describes the already proposed secure group communication protocol (SGCP) (Bhatia et al., 2013a). In SGC protocol, we considered a centralized key management center (KMC). Each user needs to submit an image IM to the KMC for registration in a grid group. KMC generates unique IDs and computes unique passwords for the users using the images submitted by the users. Each user’s password is embedded in user’s image and stego image (embedded password) is transferred to user. User uses this login information for authentication. After the completion of authentication process, KMC computes the group key and send it every member of the group for secure communication.
The computation of group key involves summation of soft dipole representation value SI of every image IM. Soft dipole representation of an image IM is a function that uniquely represents that image[Chubb C. et al., 2002]. A soft dipole representation of an input image IM is a triple SI(d, ??, ??), where d is an integer-valued displacement and ?? and ?? are pixel intensities [Bhatia et al., 2013b]. The soft dipole representation of any one-dimensional image IM with N pixels can be calculated using equation (1), the displacement d ranges from 0 to N-1(N : total no. of pixels), so any value of the d can be chosen[Bhatia et al., 2013b].
N-1 N-1-d
SI = ”” ” (”'(I[r] )(I[r+ d]))’ (1)
d=0 r=0
the displacement d ranges from 0 to N-1(N:total no. of pixels), so any value of the d can be chosen.
4. Proposed secure group message transferring stegosystem(SGMS)
This section proposes a new secure group message transferring stegosystem(SGMS) which is an extension of our previously proposed secure group communication protocol(SGCP) explained in Section 3. The main purpose of a message transferring system is to provide a secure communication channel to a sender for transferring messages to its intended receivers. In the group message transferring system, sender can share secret information with its group members by hiding it from other group members. Then, only group members of the sender’s group can sense the presence of the message and can extract the message. The other grid users cannot acquire any important information from the forward message. Figure 1 shows the proposed secure group message transferring stegosystem.
Figure 1: SGMS architecture: Architecture of secure group message transferring stegosystem
4.1 Model of the proposed system
Table 1: Notations used:
It consists of following phases
(i) Embedding group key: In this phase group key generated by KMC using SGC protocol is embedded into the images of the group members of the corresponding group. The group key is embedded using DCT steganographic method.
(ii) Transferring group key: KMC transfer the stego-images (stmg) carrying group key to all the authenticated users.
(iii) Message embedding: When any user Ui wants to transfer a message M to its group members gk : {U1, U2, ..,Um} :
Algo 4.1: Embedding
”””””””””’……………………………………………….
Input: cover image, stego key(i.e. group key), Message, iterations(for encryption)
Output: Encrypted-stego image
”””””””””’……………………………………………….
Step1. Ui embeds its message M into an image IMi using LSB with pseudorandom locations steganographic method. Embedding message bits in to the continuous pixel locations of the cover image can lead to serious security problems. To overcome this problem, message bits are embedded into the unique random pixel locations generated by the ‘Random locations generator algorithm’ (Table 2) using group key as stego key.
Step2. After embedding message, Ui encrypts the stego-image (SIM) using standard map and logistic map image encryption technique.
Step3. Ui transfer encrypted-stego image ESIM to its group members {U1, U2, ..,Um}.
(iv) Message extraction: To access the message hidden in image ESIM, users {U1, U2, ..,Um}
Algo 4.2: Extraction
”””””””””’……………………………………………….
Input: Encrypted-stego image, stego key, iterations (for encryption)
Output: original message
”””””””””’……………………………………………….
Step1. decrypts the image ESIM using image decryption method, resulting in stego-image SIM.
Step2. generate random locations using stego-key.
Step3. extract message from SIM using already shared extraction method.
Table 2: Pseudorandom locations generator algorithm[19]
Figure 2: SGMS communication model1: shows the communication steps between the sender and the receivers.
4.2 Secure message communication between Grid domains:
In above proposed protocol groups are formed based on the users working on the same application. Grid environment allows users to change its group and join/leave any group. In our protocol, KMC is responsible for regenerating new group key when a new member joins or an old member leaves the group.
Figure 3: SGMS communication model 2: steps of communication when a user joins/leaves a group
Algo 4.3. Group Key updation: User Ui of group gj wants to join the group gk
”””””””””’……………………………………………….
Input: User ID of Ui , gj, gk
Output: updated group keys of gj and gk i.e. GK_newj , GK_newk
”””””””””’……………………………………………….
Step1. Ui sends its request to KMC to leave group gj and join group gk with its user ID.
Step2. KMC verifies the identity of the user Ui.
Step3. After verification, KMC forwards its join request to members of group gk.
Step4. KMC intimates Ui on acceptance of join request from members of group gk, Ui then leaves group gj and joins group gk.
Step5. When a new member joins or an old member leaves the group, KMC regenerates the group key.
Step5.1. Updated group keys for gj and gk are:
The regeneration of group key of any group gi involves re-computation of soft dipole representation value SI again for every user’s image IMi of that group by changing the displacement factor d. Here, we changed the value of d randomly for every image using random sequence generator function. This gives unique value of SI for same image also. For computation of new group key GK_new, summation is carried out on new SI values of images of all the group members using equation no. 3.1 or 3.2 generating GK_updated .
If Ui left the group gj and joined group gk . KMC regenerates the group keys for groups gj and group gk as :
Step5.2. GK_updated[gj]= ‘(SI(i)) where m=(1,2,’tj )’Ui (group members for the
i??m Jth group leaving Ui (2.1)
GK_newj=XNOR(GK_updated[gj],Loc[i](random[rand,IMi])) (2.2)
Step5.3. GK_updated[gk]= ‘(SI(i)) where n=(1,2,’tj )+Ui group members for
i??n the Kth group including Ui (3.1)
GK_newk=XOR(GK_updated[gk],(Loc[i](random[rand,IM])) (3.2)
Where i=1 ‘. Length(IM)
Step6. KMC embeds the GK_newi into images of group members of ith group and GK_newj into images of group members of jth group.
Step7. Transfer stego-images stmg hiding new group key to the members of the ith and jth groups.
Our proposed protocol supports dynamic join/leave of the group members. To enables dynamism in the groups it provides facility of updating of group keys on change in the number of group members.
5. Security of the proposed protocol (SGMS)
5.1 Security features:
A new generated group key should be known only to the present group members. The proposed method of new group key generation fulfills the important security requirements of group key:
1. Group Key Secrecy: Even if the group key is changed q number of times, such as GK= {GK1, GK2, ‘, GKq}, it is computationally infeasible for unauthorized group member to compute any group key GKi because computation of group key involves secret passwords of all the group members. The secret password of individual user is known to that user only. The group key computed using proposed protocol is secure against brute force attack. The brute force attack on the group key is handled by increasing the keysize of the group key. The keysize of the group key depends upon the keysize of the users’ secret passwords that further depends upon the value of the displacement factor d ranges from 0 to N-1, as shown in equation 1. Here, in our proposed protocol we are choosing the unique value of d generated through random sequence generator, every time when we are updating the group key.
2. Forward Secrecy: means that new members will not be able to compute the already used old group keys. Our protocol ensures that it is computationally infeasible for any new group member to compute any old group key from the set of q old group keys such as GK={GK1, GK2, ‘, GKq }.When a new member joins/leaves a group, group key is updated for that group. Computation of new group involves generating new SI value for the images of all the group members of that group. Calculations of SI values of users’ images depend upon the unique value of displacement factor d, generated by KMC using random sequence generator. This value is XORED with unique pixel location computed from the image of user Ui using algorithm 1 to generate new group key. So it is mathematically impracticable for any group member to compute new group key.
3. Backward Secrecy: Protocol ensures that any former group member should not be able to compute new group key for group with new set of group members. It is computationally infeasible for any user to know values of displacement factor d and compute the SI values for all the present members of the group to compute GK_updated. The GK_updated is further XNORED with unique pixel location computed from the image of user Ui using algorithm 1 to generate new group key.
The attacker could not derive the group key because it is two-level secured
(i) It is computationally infeasible to generate group key as it is based on the passwords of group members and it is computed by trusted KMC.
(ii) The group key transferred to the group members is hidden into the images in such a way that attacker could not suspects the existence of it.
5.2 Security Analysis
Our protocol can handle group key modification attack. Since group key broadcasted to the group members is embedded into an image through steganography and it is infeasible for a human eye to detect the presence of information hidden in an image.
The proposed protocol is also secure from message modification attack with the use of two level security systems for hiding messages. First, the message is embedded into image using secure steganography technique and the embedding locations of the message bits depend upon the group key. Second, the stego image created after embedding message is encrypted using standard map and logistic map image encryption method.
The proposed protocol is secured from active attackers. An active attacker not able to extract the secret information can try to destroy it by applying image processing techniques. This can be avoided by using robust steganography technique. Here, we used robust steganographic system and stored group key in the discrete cosine transform coefficients of an image.
To protect the information from malicious attackers, the embedding method needs to be dependent on some secret key shared by sender and receiver. In our proposed protocol message bits are embedded into the image at pixel positions generated by the algorithm 1. This algorithm uses group key as secret key to generate the message embedding positions and only the group members can access the group key and message. So, proposed protocol is secured from malicious attackers.
6. Simulation Results
For simulating the above proposed protocol, it is divided into following phases: (i) embedding group keys in users images (ii) broadcasting stego-images(stmg) with group keys (iii) extraction of group keys from stego-images by group users (iv) embedding message in image using group keys (v) encrypting stego-image(SIM) with message (vi) broadcasting encrypted image(ESIM) to group users.
6.1. Simulation of Proposed Protocol SGMS
The experiments are done considering five groups in grid environment with different number of users in each group. The simulation is done using java programming and Matlab. We used java programs for computing group keys with different no. of users. Matlab is used for embedding group keys, messages in the users’ images and extracting group keys, messages from stego images(stmg, SIM). Matlab is also used for encryption of stego images SIM and decryption of encrypted images ESIM. It is assumed that the method used for decryption of encrypted image ESIM and extraction of information from the stego image(stmg) is already shared with the group members physically. The algorithm is tested on number of images of different images formats such as .jpg, .tif, .jpg and .png. The metric PSNR is used to measure the quality of the stego-image. The Peak-Signal-to-Noise Ratio (PSNR) is used to evaluate the visual quality of stego- images generated with the simulated method. PSNR is defined as follows:
PSNR=10?? log”10 ‘255’^2/MSE’ dB (4)
MSE = (MSE_R+ MSE_(G )+ MSE_B)/3 (5)
‘MSE_R= ‘_”i=0″ ^”M-1 ” ”_”j=0″ ^”N-1″ ‘”l ” ‘_ (X_r (i,j)- ‘I”_r (i,j)2 ) (6)
where Xr(i, j) represent the pixel values on the original image and I’r(i, j) signify the pixel values of the stego image located at (i, j) respectively. M and N represent height and width of the images. PSNR value is calculated to measure the similarity between cover and stego images, images with PSNR value greater than 30 dB is considered to be similar. In this case, it is hard to distinguish stego image from its corresponding cover image through human eyes. The PSNR of cover and stego image with embedded group key are shown in the Table 6. Table 7 gives the PSNR of cover and stego image with embedded message for proposed algorithm.
6.2 Experimental Results:
Table 3 shows the group keys of five groups with different no. of group members. Table 4 shows the experimental results for the first three phases of the proposed protocol. Table 5 gives the experimental results for the next three phases of the proposed protocol.
Table 3: group keys of groups with different no. of users
Table 4: A user Ui left group g3 and moves to group g5
Table 5: A user Ui left group g4 and moves to group g2
Table 6: Group IDs, Cover image(IM) , Group Key to be embedded, Stego image (SIM), PSNR of IM and SIM
Table 7: Group IDs, Cover image(IM), message to be embedded, Stego image(SIM), PSNR of IM and SIM, Encrypted image(ESIM).
Conclusion:
This paper proposes a protocol for secure message transferring in a group between the sender and the members of its group. Most of the researchers have used cryptography to provide secure group communication in grid environment. To overcome the weakness of cryptography that encrypted messages no matter how strong are arouse the suspicion, here, we used image steganography for hiding the group key and messages and transferring it securely to the group members. The protocol is secure against session key modification attack and message modification attack; message broadcasted to the group members is dual secured. To enhance the security of the communicated message, protocol embeds the message into image and further encrypts the stego-image. Receivers on the other end receive the encrypted image. The proposed protocol is suitable for secure message communication in grid environment. The protocol is simulated using Java programming language and Matlab.
References
Bhatia, M., Muttoo, S.K. and Bhatia, M.P.S. (2012) ‘Steganography based secure communication ‘, In Proceedings of the Second International Springer Conference on Soft Computing for Problem Solving (SocProS 2012), India.
Bhatia, M., Muttoo, S.K. and Bhatia, M.P.S 2011) ‘Secure group communication protocol’, International Journal of Advanced Engineering Sciences and Technologies, Vol.11 No. 1, pp.221-225.
Bhatia, M., Muttoo, S.K. and Bhatia, M.P.S.(2013b) ‘Secure Requirement Prioritized Grid Scheduling Model’, International Journal of Network Security, Vol. 15 No. 6, pp.478-483..
Bhatia, M., Muttoo, S. K. and Bhatia, M.P.S.(May 2013a) ‘Secure Group Communication with Hidden Group Key’, Information Security Jounal: A Global Perspective, Vol. 22 No. 1, pp. 21-34.
Burmester, M. and Desmedt, Y.G. (1994) ‘A Secure and Efficient Conference Key Distribution System’, In Proceedings. of Eurocrypt ’94 Workshop Advances in Cryptology, pp. 275-286.
Chubb C. & Yellott J. I.(2002).Dipole statistics of discrete finite images: two visually motivated representation theorems. ,Journal of Optical Society of America A, May 2002.
Cody,E., Sharman, R., Rao, and Upadhyaya, S.(2008) ‘ Security in grid computing: A review and synthesis. Decision Support System’, vol. 44, no. 4, pp. 749-764.
Foster, I., Kesselman, C.,Tsudik, G. and Tuecke, S. (1998) ‘A security architecture for computational grids’, In proceedings of fifth ACM Conference on Computers and Security, pp. 83-91.
Huang, K. and Zhang, D. (2010) ‘DHT-based lightweight broadcast algorithms in large scale computing infrastructures’, Future Generation of Computer Systems, Vol. 26 No.3, pp.291’303.
Johnson N. F. and Katzenbeisser Stefan C.(2000) ‘A survey of steganographic techniques’. Artech House Books.
Li, Y., Jin, H., Zou, D., Liu, S. and Han, Z. (2007) ‘A scalable service scheme for secure group communication in grid’, In Proceedings of 31st Annual International Computer Software and Applications Conference (COMPSAC 2007), pp. 31 – 38.
Li, Y., Xu, X., Wan, J., Jin, H. and Han, Z. (2008) ‘An authenticated encryption mechanism for secure group communication in grid’, In Proceedings of the International Conference on Internet Computing in Science and Engineering, USA, pp. 298-05.
Liao, L. and Manulis, M.(2007) ‘Tree-based group key agreement framework for mobile adhoc networks’, Future Generation of Computer Systems, Vol. 23 No. 6,pp. 787’803.
Liaw, H. (1999) ‘Broadcasting cryptosystem in computer networks’, Computers & Mathematics with Applications, Vol. 37, pp. 85’87.
Martinelli, F. and Mori, P. (2010) ‘On usage control for GRID systems’, Future Generation of Computer Systems, Vol. 26 No. 7, pp. 1032’1042.
Masque, J. and Peinado, A.(2006) ‘Cryptanalysis of improved Liaw’s broadcasting cryptosystem’, Journal of Information Science and Engineering, Vol. 22, pp. 391’399.
Park, H., Yi, W. S., and Lee, G. (2010) ‘Simple ID-based key distribution scheme’, In Proceedings of Fifth International Conference on Internet and Web Applications and Services, pp. 369’373.
Perez, J., Bernabe, J., Calero, J., Clemente, F., Perez, G. and Skarmeta, A.(2011) ‘ Semantic-based authorization architecture for Grid’, Future Generation of Computer Systems, Vol. 27 No. 1, pp. 40’55.
Smith, M., Schmidt, M., Fallenbeck, N., Dornemann, T., Schridde, C. and Freisleben, B. (2009) ‘Secure on-demand grid computing’, Future Generation of Computer Systems, Vol. 25 No.3, pp.315’325.
Tseng, Y. and Jan, J. (2001) ‘Cryptanalysis of Liaw’s broadcasting cryptosystem’, Computers & Mathematics with Applications, Vol. 41, pp.1575’1578.
Zhu, W. and Wu, C.(2008) ‘Security of the redefined Liaw’s broadcasting cryptosystem’, Computers & Mathematics with Applications, Vol. 56, pp. 1665’1667.
Zou, D., Zheng, W., Long, J., Jin, H. and Chen, V. (2010) ‘Constructing trusted virtual execution environment in P2P grids’, Future Generation of Computer Systems, Vol. 26 No. 5, pp.769’775.
Zoua, X., Dai. Y.S. and Rana, X. (July,2007) ‘Dual-Level Key Management for secure grid Communication in dynamic and hierarchical groups’, Future Generation Computer Systems, Science Direct, Vol. 23 No.6, pp.776-786.
List of Tables
Table 1: Notations used
Ui
ith user
IM
Image
GK
group key
GK_new Updated group key on member join/leave
Stmg stego image with embedded group key
M
Message
SIM stego image with embedded message
ESIM encrypted stego image SIM
gi ith group
Table 2: Random locations generator algorithm[19]
Random locations generator: random[GK,IM]
{
j[1]=10
For i=1 to m
{
a=j[i] div u
b=j[i] mod u
a=(a+hk1(b)) mod v
b=(b+hk2(a)) mod u
a=(a+hk3(b)) mod v
loc[i]=au+b
j[i+1]=j[i]+5
Print loc[i]
}
} Notations used in algorithm:
length(IM) : total no of pixels in an image IM
j: j is any value between 1′. Length(IM)
u,v: length(IM) can be represented as length(IM)= u*v
m: total no of message bits
Gk: group key
GK is divided into three keys k1, k2, k3
hk1(x) : hash function value of x using key k1
hk1(x)=x%k1
Table 3: group keys of groups with different no. of users
Group ID No. of users Group key
g1 20 984969
g2 50 405487
g3 40 275923
g4 100 271098
g5 200 342226
Table 4: A user Ui left group g3 and moves to group g5
Group ID No. of users Group key updated group key
g1 20 984969
g2 50 405487
g3 39 275923 427689
g4 100 271098
g5 201 342226 267109
Table 5: A user Ui left group g4 and moves to group g2
Group ID No. of users Group key updated group key
g1 20 984969
g2 51 405487 437665
g3 39 275923 427689
g4 99 271098 166207
g5 201 342226 267109
Table 6: Group IDs, Cover image(IM) , Group Key to be embedded, Stego image (SIM), PSNR of IM and SIM
Group ID Image(IM) Group Key Stego-image(SIM) PSNR
g1 984969 57
g2 405487 35
g3 275923 55
g4 271098 30
g5
342226
46
Table 7: Group IDs, Cover image(IM), message to be embedded, Stego image(SIM), PSNR of IM and SIM, Encrypted image(ESIM).
Group ID Image(IM) Message(M) Stego-image(SIM) PSNR Encrypted Image(ESIM)
G1 After verification, KMC 45
G1 KMC verifies the identity of the user Ui 35
G2 The experiments are done 43
G2 the file is embedded 48
G3
Grid environment allows users can change its group
37
G3 Our proposed protocol supports dynamic join/leave 41
G4 Join these two files 38
G4 Enter the data in these two files 44
G5 Group key is updated 35
G5 File is encrypted 48