Implementation of Security in Cloud Systems Based using Encryption and Steganography

Implementation of Security in Cloud Systems Based using Encryption and Steganography 1 A.Mahesh Babu, 2 G.A. Ramachandra, 3 M.Suresh Babu 1,2 Department of Computer Science & Technology, Sri Krishnadevaraya University, Anantapur. AP, India. 3 Department of Computer Science, Aurora s Technological and Research Institute, Hyderabad Abstract The use of cloud computing has increased rapidly in many organizations. Security is considered to be the most critical aspects in a cloud computing environment due to the sensitive information stored in the cloud for users. The goal of cloud security is mainly focused on the issues related to the data security and privacy aspects in cloud computing. Researchers have discussed many security techniques in cloud computing systems. However, since there are many malicious people who want to access secret and personal information stored in cloud servers, information security is still a serious problem to solve. To conceal and protect secret and personal information, in this paper we propose and develop a system that can encrypt secret message and embed it into an image file by using a new morphing based steganograph technique. Even if the malicious people may steal the image containing the secret information, they cannot read the secret information because this is almost impossible without the stego key. Consequently, improvement of security in cloud systems by using cloud systems based is implemented. Keywords : Encryption, Steganography, Stego Key I. INTRODUCTION In recent years, many cloud computing systems have been developed, and have been applied not only to scientific computations, but also to document sharing, web-based mail service and so on. Among them, onlinestorage service is a typical utility form of cloud computing system. Basically, each registered user has his/her own independent directory, and any user cannot access other users directories without permission. However, in some cases, files saved in cloud servers can be read and written by some third party who knows cyberattack techniques well. This is a big security issue in cloud computing. If we do not solve this problem, cloud computing may cause a tremendous social problem in near future. In this paper, we propose a prototype cloud computing system that employs a new steganograph technique for improvement of security. Our system encrypts the secret message by hiding it in a synthesized image that can be generated dynamically through image morphing. If some malicious attackers may obtain the image containing the secret message (i.e. the stego-image) via the transmission channel or server, they cannot access the entity of the information (the message) because the secret is protected by a huge Note that this kind of protection methods, which uses the steganographic image, have already been developed before. However, our system differs from the conventional ones. In our system, as we mentioned, we use the dynamically generated morphing image file for. This file is used both for embedding the message and for encrypt the message. The purpose of this paper is to build a secure prototype cloud system. This is the first step in our study. For further improvement, we will use machine learning and computer vision techniques for automatic image morphing. Using results obtained in machine learning and computer vision, we can generate cover images using any images selected or uploaded by users. It means that sets of reference images will increase, and the combinations of reference images can be increased. Therefore, we expect that this can farther improve the security. The remainder of this paper is structured as follows. Section 1.1 provides fundamental knowledge in this paper. Section 2 explains our method. Section 3 shows the design of the proposed system. Section 4 discusses results and possible problems. Finally, section 5 describes conclusion and suggests future work. 1.1. FUNDAMENTAL KNOWLEDGE In this study, we use two image processing techniques, namely image morphing and image based. These two techniques will be used to improve the security in cloud computing environment. A. Cloud Computing System Cloud computing system is a kind of utility form of computer resources via network. It is defined as follows by National Institute of Standards and Technology (NIST). [1] Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model is composed of five essential characteristics, three service models, and four deployment models. Currently, there are various kinds of utility forms for cloud computing. For example, to store images uploaded from mobile devices to the server; to edit the document files on web-based application; and so on. The common point is to use stego-key. Thus, the proposed cloud system is secure. 80

computer resources via network, and users just pay some charges based on the services. B. Image Morphing Technique Image morphing is a technique of computer graphics. Its main purpose is to make a natural (neutral) image from two or more (reference) images. [2] Usually, image morphing provides synthesized natural images based on some feature points defined in the reference images. In our method, we use this technique with some of the images taken from a given reference image set and some of random parameters as the weight, of the images, and generate a natural morphing image. We use this generated morphing image as a key for encoding and decoding the secret message. C. Steganographic Technique Steganographic technique () is one of the techniques for information hiding. [4] [5] In this technique, some messages or data are hidden in multimedia datum. The multimedia datum, which is used in this technique, has redundancy, and therefore if we modify it slightly, we may not be aware of this difference. There is one of big difference between steganographic technique and encryption technology. First of all, the main purpose of encryption technology is to prevent datum from decryption. On the other hand, the main purpose of steganographic technique is to prevent the existence of the secret message from being detected by third party. That is, if the third party does not know where is the secret, he or she will not be able to attack the secret. This is extremely important in a cloud computing environment. II. METHOD In this section, we explain the methods of image morphing and generation of steganographic image. A. Image Morphing For convenience, we use the definitions given in Table 1. Symbol Description Defintion RI Reference image set {I 1,I 2.I n } N Size of reference nϵ N, n 2 image set SI Source image set SI RI. SI = m M Size of source image mϵ N, 2 m n set I i Reference images I i ϵ RI w i Weight for each 2 i m, 0 w i 1 source image P Accuracy of weight p ϵ Q In image morphing, we use reference image set RI and weight value for each source image wi. We generate the morphing image by the following steps. Step 1 : Create a source image set SI from RI. SI should have m images and it should have no duplicates. Step 2 : Generate m random values wi. Step 3 : Decide the accuracy of weight. Omit wi smaller than 10 p (The number of non-zero weights is larger than or equals to 2). Step 4 : Generate a morphing image with SI and wi. By using these parameters, Sizepattern = 10( ) possible morphing images can be generated. Here, we show an example. We set the size of each set as follows. RI = 1000 SI = 5 n = 1000 m = 5 p = 10 By using these parameter values, the number of possible morphing images can be calculated by 1050 (1000 5 ) 2200. Therefore, the strength(encryption strength) is about 200 bits Figures 1 and 2 shows the flowchart of message hiding and message extraction. III. DESIGN OF SYSTEM In this paper, we construct the cloud computing system in local network, and investigate about safety, security and so on. From the hardware side, this cloud system is constructed with two servers and a number of ipad. This kind of system, which constructed some of servers and a number of mobile devices, is usually used in many services. On this system, we implemented server-side and client-side applications. 81

storage server Create the stego-image based the message and the cover image Decode message from the stego-image. A. System Configuration In this study, we constructed a system with two servers and a number of clients (ipad). The purpose of each device is shown in Table II. B. Implementation In this section, we explain the implementation of each component. 1) Server C: In the cover data generator (Server C), we implemented the system on Linux system. Currently, the system provides only user authentication and morphing image generation for received requests. On this server, we used GNU Octave for carrying out high-speed morphing image generation function. In this phase, before the morphing generation, user authentication is required. The flowchart of Server C is shown in Fig. 3. 2) Server S: In the storage server, we implemented the system on Linux system as for the cover data generator. The system provides user authentication and file upload/download function.. 3) Client *: For each client, we implemented the system on ios as end-user client. The GUI of the implemented application is shown. In this application, we implemented the following items: User authentication for each server; Create the request of cover image generation to server C; Receive the cover image from server C; Upload/Download the stego-image to/from server S Table II : Definitions of components in the Prototype system Name Abbreviation Role name Cover data Server C Dynamic generation generator of morphing image Storage server Server S Store the steganographic image which is uploaded from client Clients Client * Input the message, hide it into morphing image and upload it to the Fig 3 : Implementation of AES algorithm with Fig 4: Implementation of DES algorithm with Fig 5: Implementation of Blow fish algorithm with Fig 6 : Implementation of RSA algorithm with 82

Fig 7 : Implementation of Random algorithm with Algorithm Resolution Size (Before) Size (After) Pixel count Execution time (Before) Execution time (After) ticks AES 600 x 900 289060 70170 54000 7 72108 DES 600 x 900 289060 70201 54000 4 49540 Blowfish 600 x 900 832406 832406 54000 3 35379 RSA 600 x 900 289060 70234 54000 175 173514 Random 600 x 900 289060 70201 54000 3 35266 Steganography 600 x 900 289060 70216 54000 0 5481 C. System Requirements We restricts the operating environment with the aim of improvement of security. The restriction and advantages are shown below. Desired operating environment is shown in Fig. 6. 1) The cover data generator (Server C) should be accessed from only the private network. By applying this restriction, we can prevent the cover images (which are also used as the stego-key) from leaking to the public network (or WWW). If there are no keys, the third person cannot decode the stego-image, and the system can be secure. 2) Server C and Server S should be independent of each other. By applying this restriction, if either server system was under attack, another system s safety can be guaranteed as much as possible. As the reason for that, we need to obtain some of data from these two of servers for decoding. Moreover, if we decode the stego-image, we need to find out the relation of the data in each server. Fig. 8 shows the desired operating environment. D. Comparison Here, we compare the proposed system with the conventional one. In conventional systems, they uses HTTPS (HTTP over SSL/TLS) for data transmission in place of normal HTTP. HTTPS is a kind of URI scheme, and it means encrypted HTTP connection with SSL (Secure Socket Layer) or TLS (Transport Layer Security). However, some of security issues of TLS are Table 3 : Shows performance of various algorithms. reported recently. Some of unknown problem may possibly hide even though this problem was already fixed by developers in the world. In other words, if the attacker, who wants to obtain the information, found the unknown problem in this kind of encryption technology, the information can be stolen. This is a very big problem in (secret) data transmission. In the proposed method, we use techniques before data transmission instead of data encryption (in fact, the data are encrypted before applying ), and this is the major difference between our system and conventional ones. In our system, we apply the techniques to data which will be sent. After this process, the morphing image covers the data. In data transmission, even if the attackers may successfully attack and obtain the stego-images, they cannot perceive if the stego-images really contain secret data or not. In the worst case, even if the attackers detected the existence of the secrets, they cannot decode the data because the stego-key is not available from the transmission network. Because, the key is the cover image, which is obtained through morphing based on some unknown reference images, and the cover image itself is not available from the public transmission network due to the restriction described previously. The cover image is generated dynamically and it is sent in private network, so it does not exist in global network, unless otherwise someone who can access this private network leaked this image. IV. RESULTS AND DISCUSSION Here, we discuss the security of our proposed method. In our proposed method, there are some advantages. 1) The attackers cannot perceive that the stego-image has the message internally. Because, the stego-image is natural and basically human can not perceive the difference between the stego-image and natural image. 2) Even if the attackers can obtain the stego-image, they cannot 83

decode it. Because, in global (public) network, there is no cover image which is the key for encoding and decoding. Therefore, if the attackers cannot obtain the cover image from private network, they cannot decode the data hidden in the stego image. 3) Even if the attackers can obtain the stego-image and access the cover image generator (Server C), it is nearly impossible to decode the image. Almost of all reasons are same as last explanation. In this case, the attackers can access Server C, so they can get the key from the server. However, there are no relations are described in stegoimage, so if the attackers want to get the cover image as a key for a given stego-image, they must try all possible cases for generating the image. If they try this, it can take a very long time for calculation (refer section III-A). In other words, we have the time to detect this as attack. V. CONCLUSION In this paper, we proposed a way for improvement of security in cloud system based on. In the proposed system, we used the dynamically generated morphing image for improvement of security, and this image covers the message which we want to hide. This is application of steganographic technique, and its image is natural, so human cannot perceive the image hides the message internally. Moreover, this morphing image is generated dynamically, and there are no keys for decoding it. Besides this technique, we defined some rules of operating environment for improvement of security. In the future, we should apply machine learning to morphing technique which we used in this paper. We expect that application of machine learning may improve the security further for two reasons. First, machine learning can be used for automatic morphing, so that a large number of morphing images can be generated effectively. Second, machine learning can find feature points good for morphing for any given images, and thus the reference image set can be virtually infinite, and the security strength will be very high. We expected some attacks to our implemented system, and we discussed how to cope with it in section V. In next step, we will verify the quality of security of our implemented system. REFERENCES [1] National Institute of Standards and Technology, Peter Mell and Timothy Grance, 2011, The NIST Definition of Cloud Computing, pp.2. [2] G. Wolberg, Recent advances in image morphing, Proc. Computer Graphics International (CGI96), pp. 64-71, 1996. [3] M. Steyvers, Morphing techniques for manipulating face images, Behavior Research Methods, Instruments, and Computers, Vol. 31, No.2, pp. 359-369, 1999. [4] S. Kondo and Q. F. Zhao, A novel steganographic technique based on image morphing, Proc. International Conference on Ubiquitous Intelligence and Computing (UIC06), Wuhan and Three Gorges, China, pp. 806-815, Sept. 2006 (Lecture Notes in Computer Science 4159, Springer). [5] S. Katzenbeisser and F. A. P. Peticolas, Information hiding: techniques for and digital watermarking, Artech House, 2000. [6] Jijo.S. Nair and BholaNath Roy, Data Security in Cloud, International Journal of Computational Engineering Research(IJCER) (ISSN: 2250-3005), pp. 78-79, 2012. [7] WAWGE P.U. AND RATHOD A.R., CLOUD COMPUTING SECURITY WITH STEGANOGRAPHY AND CRYPTOGHRAPY AES ALGORTHM TECHNOLOGY, World Research Journal of Computer Architecture Volume 1, Issue 1, 2012, pp.-11-15. 84