A Hybrid Soft Computing Approach on Security Algorithm of Cryptography

Abstract

Cryptography is a security mechanism to secure data from unauthorized access during the data transmission over the public channels. Ubiquitous use of internet over public channel makes easy to data retrieval from data transmission. Many algorithms are already exited as such RSA, DES, BLOWFISH and many more. In this paper we will proposed and implement an effective cryptography security method based on soft -computing techniques (neural network and Genetic). So the result is going to be additional correct then previously implemented methods. In depth experiments demonstrate that our new proposed method achieves fascinating performance on real-life events over other methods. Proposed system our main aim to find an easy and fast solution for users to send data with a more secure and trusted way over the public channels. So that their personal information not captured by any unauthorized entity. Keywords: Cryptography, Symmetric key, Asymmetric key, RSA, DES, Neural Network. I. INTRODUCTIONSecurity is a simple need in human environment in all perspective. These days the vast majority conveys over open channel and there, they transmit data that is extremely delicate and individual. So plainly security is extremely and essential worry in present time. Cryptography is a route by which we can secure our information from unapproved get to or illicit passage of obscure substance. In cryptography we utilize some numerical technique that progressions the first information so it cannot be perused without a key. The primary target of this paper is to applying the neural system in cryptography for secure information correspondence and to enhance the proficiency of cryptographic calculation. Essentially neural system used to take care of the intricate issue like: predication, self association, self learning, irregular number and so forth. The principle qualities of neural system are system structure, adaptation to non-critical failure, aggregate arrangement, learning capacity, parallel handling, and conveyed memory. Here with the utilization of neural system we propose another successful cryptographic calculation for secure information correspondence.