|
|
 |
| |
|
|
|
Automated Classification of The Sybil Attack in BlockChain Network Using Multi Neural Memory Network Classifier |
|
|
|
PP: 323-336 |
|
|
|
doi:10.18576/amis/170214
|
|
|
|
Author(s) |
|
|
|
D. Nancy Kirupanithi,
A. Antonidoss,
G. Subathra,
|
|
|
|
Abstract |
|
|
| Block chain technology is a distributed form of digital record that does not need any central authority. As all the information available in the block chain are transparent, everyone can see it. Since blockchain technology is almost hack-proof, it has attracted several industries like business, academia, government, and healthcare. Common concerns regarding Blockchain technology’s impact on privacy and security are unfounded, and may be attacked in a number of ways. We discovered that the Sybil attack has a significant effect on public/permissionless blockchains because an attacker may undermine the blockchain by generating a large number of pseudonymous identities (i.e. Fake user accounts), making genuine entities a minority. These artificial nodes may mimic real ones and have outsized effect on the network just as they would if theywere real. This might result in a cascade of assaults including denial- of-service and distributed denial-of-service. In this work, we show how a Sybil attack might reduce the throughput of a blockchain test bed. We provide a solution directive in which every node keeps tabs on the actions of the others, looking for the ones that are forwarding the blocks for only one user. Initially by preprocessing the input data from network are normalized. Then by smart contract the transactions are created. By promptly identifying, blacklisting, and notifying other such nodes in the transactions, the Sybil attack’s reach may be restricted by using the Multi Neural Memory Network (MNMN) classifier. Then the data can be securely stored in the block chain ledger by using the stream cipher Crypto Fish algorithm. We analyze experimental results of the proposed solution under simulation environment. The evaluation of performance is observed and compared with other traditional method. The proposed model is proved to be an efficient one in its throughput, accuracy and execution time etc. |
|
|
|
|
|
|
|
