Abstract | ||
---|---|---|
As a representative architecture of content- centric paradigms for the future Internet, named data networking (NDN) enables consumers to retrieve content duplicates from either the original server or intermediate routers. Each node of NDN is equipped with cache that buffers but not validates the data, making it vulnerable to various attacks. Cache pollution, one of the specific attacks in NDN, fraudulently alters the cached contents by excessively requesting worthless information, squeezing the space of real popular contents and thus degrading the experience of normal users. In order to address the issue, this paper proposes a defense scheme based on deep reinforcement learning (DRL) against cache pollution attack, in which whether a data packet is to be cached is decided by a trained intelligent agent, that is adaptive to dynamic network states and following long term rewards, the accumulative data-requesting delays. Finally, the DRL-based scheme is evaluated and compared to two other existing schemes. Experimental results show that the proposed defense mechanism outperforms the others significantly, and is proved to be effective against cache pollution attacks. |
Year | DOI | Venue |
---|---|---|
2021 | 10.23919/JCIN.2021.9387728 | Journal of Communications and Information Networks |
Keywords | DocType | Volume |
named data networking,cache pollution attack,deep reinforcement learning | Journal | 6 |
Issue | ISSN | Citations |
1 | 2096-1081 | 0 |
PageRank | References | Authors |
0.34 | 0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jie Zhou | 1 | 2103 | 190.17 |
Jiangtao Luo | 2 | 11 | 4.23 |
Junxia Wang | 3 | 2 | 2.45 |
Lianglang Deng | 4 | 0 | 1.01 |