Big data encryption
Privacy and security in the context of big data are critical issues. In the case of complicated applications, the big data security paradigm is not recommended, hence it is disabled by default. However, in its absence, data can always be easily corrupted [12].
Big data may contain sensitive information about persons, privacy is a crucial consideration. To address the privacy issue, data can be de-identified by deleting traits that would allow an individual to be identified. This is a technique that, when done effectively, works both while data is controlled and when it is released [13].
Encryption is a powerful method for ensuring data security. The essence of data encryption is to use algorithms to convert the original plaintext file or data into an unreadable string of code known as ciphertext. Even if someone intercepts the distorted code, he or she cannot utilize it to obtain the original message. This efficiently preserves the secrecy of the data and prevents tampering with the data Users with access can decrypt the file using the matching private key, then update or modify the ciphertext. There are two types of encryption: symmetric encryption and asymmetric encryption. To encrypt and decrypt data, symmetric encryption employs a secret key [14].
Encryption also is the most widely discussed approach, and it can protect data secrecy and integrity. Because not all encryption methods are created equal, cloud service providers and users must employ the most recent encryption techniques (Homomorphic encryption, AES or DES) and longer keys, which strain processor capabilities [15].
Transparent data encryption concept
Shmueli and Vaisenberg investigate five traditional database encryption architectures and compare them. The writers illustrate that existing design can offer a high degree of security, have a considerable effect on performance and impose large changes on the application layer or can be transparent to the application layer and provide great performance [16].
However, the use of encryption in the database system will impact the performance of a system. As stated by Sharma and Trivedi [17], based on the research that has been done, taking an approach to modeling from several levels for error prevention and safety may require sacrifices based on several attributes such as Efficiency and Reliability. Other research conducted by Wolter and Reinecke showed that the combination of security and performance poses interesting tradeoffs and inspires similar models as the combination of performance and dependability, known as performability [18]. The consequences of security on the probability of having a particular system state require more performance, judging by the increased need for transactions for encryption time, combined performance and security (CPSM), and ongoing transactions.
Transparent data encryption may help, however solutions for TDE supplied with important database systems only ensure a data-only system, and are seemingly unnecessary if the adversary can access the computer physically, which poses a likely concern when hosting in the cloud. This work provides an alternative approach to TDE, taking into consideration cloud-specific hazards, extends encryption to cover data in use and partially information in motion and is able to run huge SQL sub-sets including heavy relationship operations, complex attribute and transaction operations [16].
Transparent encryption technology allows data to be encrypted throughout the process without altering user habits. It's an encryption algorithm that is also stressed in encryption as "transparent." The window system now potentially have a useful application with transparent encryption technologies. The hook software intercepts the opening function of the file when the user opens the file. The file is copied to the hidden directory folder, decrypted data and provided to reader to obtain a clear copy before the file information is sent to the reader. The hook application also can intercept the closing process if the user shuts the file, encrypts the file above, before saving to the storage device and then transfers it to the original folder. This completes the transparent process of the complete document encoding and decryption [19].
A study done by [20] focuses on Transparent Data Encryption, a technique that is used to tackle data security issues. Transparent encryption implies that databases are encrypted on a hard drive and on any backup medium. Today there are many security dangers and compliance problems in the global corporate world need security solutions that are transparent by design to defend against data theft and fraud. Transparent Data Encoding provides a transparent, standard-based security system that secures network, disk and media data. By transparently encrypting data it is straightforward and efficient to safeguard the stored data. High security levels for columns, tables and tables that are database files saved on hard drives or floppy drives or CDs, and other protection information.
Transparent Data Encryption (TDE) offers transparent, standard-based security for network, disk, and backup data protection. By transparent encryption of data, it is easy and efficient protection of stored data. TDE is able to encode and decrypt data and log files in real-time. The encryption employs a Database Encryption Key (DEK), which is saved for recovery in the database boot. The DEK is a symmetric key encrypted with an EKM module protected certificates in the server's master data base or with an asymmetric key. TDE secures 'rest' data, which means data and log files. It enables many rules, regulation and guidelines made in different industries to be complied with [21].
In fact, TDE works effectively, if the backup of your database to be protected. You need a master key, a certificate to restore if you are implementing TDE in the source server and wish to restorate your database to another server. Think about opening your bank locker. One key is to implement an extra layer of protection with you and the other key is the prohibition specialist. The Always Encrypted (AE) allows transparent encryption of client apps from the database. This AE function is enhanced by TDE by the addition of an enclosure layer in the memory and transit of sensitive data as well as in rest. In fact, the Always Encrypted Driver encrypts and decrypts the application. Any potential leakage to database administration can therefore be managed by the information owner by keeping the decryption keys, in order to prevent administrators from accessing sensitive data. In contrast, the database administrator uses the master key and certificates to access the TDE encryption keys [22].
Thus, determining on how much percentage of performance would be taken from using TDE and not using TDE. SQL interface encryption implemented is: (1) the sensitive table is renamed; (2) the sensitive table is encrypted, (3) the encryption trigger is defined; (4) the decryption view has been defined. In theory, the application layer should be transparent in this architecture. In practice it is not, however, as: (1) some actions cannot be executed on the view and must be reprogrammed to utilize the renaming table (for example, insert, update or truncate). (2) No questioning of range is supported. The aforementioned cache design is to be implemented in MySQL, and only by modifying two strategic areas in the InnoDB storage engine: the cache location (added decryption) and the cache value location (add encryption). Similar to the aforementioned cache architecture, the storeroom architecture implementation requires only two strategic points in an InnoDB storage device to be changed: the location where a site reads from the disk (i.e. all cells in this page will be decrypted) and the location of the site on the disk (i.e., encrypt all cells in this page) [16].
Reliability and efficiency
This research was conducted by initiating performance testing and compare the implemented TDE's SQL Server and non-implemented TDE's SQL Server. We focus on performance value such as Reliability and Efficiency to know how significant performance degradation once a system is implemented by TDE by doing Load Testing, Stress Testing and Backup Testing. Each of test can show how affected the systems are by implementing TDE on database system.
Challenged with greater hurdles than demands for results, reliability and availability. For example, failure rates for software systems are exceedingly difficult to assess unless software testing is improved [23].
Reliability, availability models, and recovery and maintenance times are as well as models that are used to drive the models, such as failure rates, recovery success rates. The state of SW systems can be very vast to limit the application of analytical models of availability. Simulation models with substantial approximations therefore become the only means for testing availability and reliability [23].
Performance testing
The result from this paper will be achieved using HammerDB as a benchmarking tool and Performance Analysis of Logs (PAL) as an analysis tool. This paper aims to provide knowledge about how Transparent Data Encryption would affect the database system's performance and how substantial the degradation of performance is by gaining a security measurement on the system.
The performance of a machine learning-based approach, particularly modern machine learning, is well recognized to depend on several exercise models. This variety can be expressed by number of subjects in a given scenario, such as our case. In this part, the impact on recognition performance of the number of training subjects is examined [24].