Secure Multicast Communication Based On Chaotic Neural Network

With the development of network communication technology，there aremore and more opportunities to transmit informations in people’s living andworking through the network or other channels. In order to save channelresources and improve the rate of informations transmission. Multicastcommunication has been brought. Although more and more people payattention to multicast，it’s development of application still be slow. The safetyof its application subject is one of the most important restrictions. It’s aneffect way of ensuring insecure multicast to encrypt multicast message.In recent years,as the development of computing technology andcryptanalysis developing,in order to ensure many traditional cryptographicalgorithms’(RSA, ECC，etc.) security，we have to continue to increase thelength of the keys，but this way will increase the computational complexityand reduce the encryption speed. Therefore one of the priorities of today’scryptography researches is seeking an highly complex and able to carry outfast parallel processing of of the encryption algorithm.This paper first introduces the multicast and the cryptography necessaryknowledge，and introduces the relationship between them and the neuralnetworks， and proposes two algorithms based on the chaotic neuralnetwork.The schemes use the parallel and chaotic feature of the neuralnetwork.Neural network in the sequence of periodic、 random、 linearcomplexity and reconfigurable encryption algorithm based on the serial shiftregister has an advantage in the password system，the rational use of neuralnetwork can improve the security of the encryption algorithm. Analysis theefficiency of algorithm and security of algorithm’s encryption anddecryption， and using vc programming algorithm， and using the testingstandards of evaluation of random sequences and the pseudo-random sequence SP800-22which made by national institute of standards and technology to test thekeystream and ciphertext. The experiments show that the random sequencehave better randomness.According to the characteristics of centralized group key managementprotocol and the cryptographic algorithms based on the chaotic neuralnetwork，this paper presents the symmetric encryption algorithm used inGroup Key Management Protocol,which can solve the scalability issues of themulticast group well, and improve security of multicast communication; andpresents the hybrid encryption algorithm based on the chaotic neural networkused in Logic Key Hierarchy Protocol,which can leverage resources to achieveefficient encryption and universal applicability.