SECURE AND EFFICIENT RETRIEVAL OF VIDEO FILE USING BLOOM FILTER AND HYBRID ENCRYPTION ALGORITHMS

Video files are the important source for the big data. Storing, managing and processing the video files are becoming more challenging day by day. Securing and retrieving the video files from the cloud storage has become one of the hot research topic as the vulnerability remains high. Usually, the confidential video files such as CCTV monitoring, OTT videos are transferred to the cloud storage using an insecure communication network. To protect the video file, existing technique uses a symmetric or single key encryption, which uses only one key to perform encryption and decryption. Symmetric key encryption reduces the work load key management and computation time. However, symmetric algorithms are vulnerable to various attacks as they use only single key to perform encryption and decryption. Likewise, existing video file encryption are inefficient to retrieve a particular block of data. In this research work, we create a hybrid model, which integrates the symmetric and asymmetric key encryption to tighten the security. Likewise, to retrieve a particular block of video file from the cloud storage, a bloom filter based index structure is maintained by the cloud service provider. The experimental result shows that the proposed model reduces the computation time and improves the security of the multimedia files stored in the cloud.


INTRODUCTION
With the fast growth of social media, online gaming, OTT platform, IoT devices and CCTV cameras, the multimedia data generation rate has increased exponentially in recent years.
According to the 'Big Data Analytics Industry Report 2020' survey, over 2.5 quintillion bytes of data are generated each day. In addition, it also estimates that the amount of digital data produces rate are daily will increase up to 3.5 quintillion bytes in 2025 [1]. Securely transferring, storing and managing the multimedia data in the cloud storage remains as a challenging and tedious tasks to the Cloud Service Providers (CSP). Recently, a video messaging app named 'Dubsmash' announced that hackers nabbed nearly 162 million users account holder names, email addresses, passwords and the video contents from their servers. Likewise, in 2017, the HBO OTT contents were hacked and around 1.5terabytes of multimedia files were illegally circulated among the internet [2]. By encrypting the user's confidential multimedia data, the vulnerability towards data alteration and unauthorized access can be stopped in the cloud environment. However existing single key encryption techniques such as AES and DES creates a large amount of cipher data upon encryption [3] [4]. As far as cloud is concerned, it allows multiple authorized user to approach the data stored in the cloud though internet. Since, the symmetric keyis shared to multiple authorized users in the cloud environment, the possibility of key leakage is high [5].
On the other hand, it is important also to reduce the computation overhead on the device while performing encryption [6], because nowadays devices with limited resources, such as, mobile phones, connected cameras and IoT devices are used as an end device to capture and transfer the data the cloud storage servers. Therefore, taking into consideration of limited resources, a video encryption algorithms need to be developed. For realworld applications, a video encryption algorithm has to take into account various parameters like security, computational efficiency, compression efficiency and so on. On the other hand, existing algorithm takes too much of time in retrieving the video files. Because, they doesn't use any index based structure to retrieve the multimedia data blocks.

RELATED WORKS
This section discuss about the several existing woks. Jagadeeshwari et al have proposed a Dynamic Bloom Filter Hashing based Cloud Data Storage (DBFH-CDS) Technique [7] for improving the security as well as confidentiality of the data storage in a cloud environment. They 5527 SECURE AND EFFICIENT RETRIEVAL OF VIDEO FILE have used the data fragmentation model for fragmenting the large cloud datasets. They have also employed Bloom Filter in DBFH-CDS Technique for storing the fragmented sensitive data increased security. Sourin Chakrabarti [8] have proposed an efficient and modified approach for image retrieval using multiple neural hash codes. They have limited the number of queries by identifying the false positives with the use of bloom filter. They have used the local deep convolutional neural network that combines the powers of both the features of lower and higher layers. They create the feature maps which are further compressed with the use of PCA. This is then fed to the bloom filter after doing the binary sequencing with the use of a modified multi kmeans approach. The feature maps obtained are additionally utilized in the image retrieval process. They first compares the images in the higher layers for finding the images that are semantically similar and gradually moves towards the lower layers for searching structural similarities. Mai Jiang et al [9] have developed an improved algorithm based on Bloom filter application for bar code processing and recognition. The bit vector of Bloom filter is divided into two parts. Every element is mapped to the part of the bit vector with the use of hash functions.
Each element is amplified and is mapped to another part of the bit vector by using the hash function. They states that their algorithm reduces the false positive rate of the Bloom filter also it does not increase the time and space costs. Thilina et al have proposed two techniques [10] [11] that can be applied on Bloom Filter encoding for improving the privacy against attacks. They use neighboring bits in a BF that generates new bit values. They have made an empirical study on large databases and are compared with the proposed techniques and states that the proposed model provides high security against privacy attacks, and achieve better similarity computation accuracy. Raghavendra et al [12] have made a survey and have investigated the various aspects of data sharing in different manner such as user revocation, encryption techniques, identity privacy, competency and key distribution. Different schemes such as Plutus, Sirius, Secure scalable data access schemes are analyzed and found that they have improved proxy encryption.
The Multi-owner DataSharing is also discussed based on the above mentioned significant parameters.

SYSTEM MODEL
The proposed model intends to create a secure framework for protecting the user's multimedia data from unauthorized users in the cloud storage servers. In the proposed work, before encrypting the data using hybrid algorithm, two important operations are performed, such as, data chunking and data hashing. Data chunking helps the data owner to split the multimedia data into several blocks and makes encryption easy. It also facilitate the user to decrypt a particular block of data upon retrieval. Likewise, data hashing is used to produce the message digest for each block and create a hash index.

Data Chunking
Data chunking is an important process in data management while outsourcing the user data. The plain text (unencrypted user data) can be of any size ranging from 1KB to 'n' TB.
Chunking is a process of breaking down the larger user data into smaller data chunks before encrypting. The proposed scheme uses fixed size data chunking mechanism to create even sized data chunks. The chuck size is fixed to 1024Kb. Algorithm 1 explains the fixed size data chunking method below,

Data hashing and creation of Bloom filter
To create the hash table in the cloud service provider, Bloom filter is used It is a spaceefficient probabilistic data structure used to test whether an element is a member of a set. In the proposed work, after splitting the multimedia data into several small-sized data blocks, 'n'

Hybrid encryption
The proposed hybrid algorithm combines the convenience of a public-key cryptosystem with the efficiency of a symmetric-key cryptosystem. To improve the security while transferring the data to the cloud Triple DES algorithm is used [15]. Likewise, to protect the data from internal and external attacks in cloud storage server, a public key cryptosystem called ECC (Elliptic Curve Cryptography) is used to encrypt the data blocks [13].Triple DES is an encryption technique which uses three instance of DES on same plain text. It uses there different types of key choosing technique in first all used keys are different and in second two keys are same and one is different and in third all keys are same. The working process of Triple DES algorithm is shown in Fig. 2.   FIGURE 2: Encrypting the data block using Triple DES Once after the message digest MDi are encrypted using Triple DES algorithm, it will produce a 168 bit cipher text. Later, these cipher text will be considered as a plain text and reencryption will be performed using ECC algorithm.
Elliptic Curve Cryptography (ECC) is a key-based technique used for encrypting data.
ECC focuses on pairs of symmetric and asymmetric keys for decryption and encryption.

Retrieving a particular data block using Bloom Filter
Reducing the time of the retrieval and allowing the authorized users to directly access a particular block of multimedia file is important. The proposed work uses a Bloom Filter based approach to retrieve the data block in a fast manner. To retrieve the particular data block, user proves the ownership and checks the corresponding hash location in the hash table. If all the bits are set to 1, then the data block is retrieved from the cloud server. Algorithm 3 explains the retrieval of particular data block from the cloud storage server

IMPLEMENTATIONS AND RESULT DISCUSSION
The proposed hybrid encryption algorithm is analyzed by executing a set of experiments.
The experiment iscarried out in a eucalyptus private cloud which includes cloud controller and walrus as storage controller. Private cloud was installed on a server with the specification of Intel Xeon processor, which has a processing speed of 2.1GHz, 64GB of RAM memory and 4TB of storage space. The test used 500 files of real video data set, uploaded into the storage anddownloaded based on the user's requirement. The experiment result clearly shows that the hybrid encryption algorithm provides secureness for the cloud users to store the data in the cloud. Results demonstrate that the hybrid encryption algorithmis highly complex in nature and the time taken for the encryption and decryption operation is reduced in a higher rate.
Performance analysis metrics are done based on the experimental setup.
To analyze the performance of the proposed technique, avalanche effect is measured and compared with other existing algorithms. Avalanche effect is a property in which the change in output with respect to the input can be measured.It will be hard to perform any kind of analysis in the cipher text. Hence avalanche effect has to be measured for observing the level of change [14]. An effective security algorithm is a one which has more avalanche effect. From     Figure 4, clearly shows, how far the proposed hybrid encryption algorithm reduces the time of the encryption. As the file size increase, the system performs faster. So the system will give predominant results against the large file, which is feasible to be implemented in cloud environment.

FIGURE.4 Time taken to perform encryption
We can observe that encryption time is reduced to a greater extent. It is also observed that the proposed technique takes too much of time to encrypt without pre-processing module. From Figure 4 and Figure 5, it is observed that the time take for encrypting as well as decrypting using proposed system is less when compared to other existing public key algorithms.

FIGURE.5 Time taken to perform decryption
The proposed pre-processing model reduces the encryption time and makes the system work faster. Even though the time taken for pre-processing the user data is high, it will highly reduce the working time of the encryption module.

CONCLUSION
Thus a hybrid model is created that integrates both the asymmetric and symmetric key encryption to tighten the security. A Bloom filter based index structure is created and maintained by the cloud service provider for retrieving the video blocks from the cloud storage. The experimental result shows that the proposed method reduces the computation time and improves the security of the video files stored in the cloud. The proposed pre-processing model reduces the encryption time and makes the system work faster. Though the time taken is high for preprocessing the user data, the working time of the encryption module is reduced much.

CONFLICT OF INTERESTS
The author(s) declare that there is no conflict of interests.