Due to the increasing demand for applications by users worldwide, many service providers (such as Google and Microsoft) have deployed their services on a Cross-Data Center Network (DCN) connected by a Wide Area Network (WAN). While WAN ensures interconnection between headquarters and nodes at all levels, security issues within the network are becoming increasingly prominent. In the event of a malicious attack, data transmission across DCNs may result in packet loss, leading to various adverse outcomes such as data transmission delays, low throughput, and wastage of network bandwidth. To ensure the reliability of cross-DCN traffic transmission in the face of malicious traffic attacks, this paper proposes a coding method designed to achieve low delay and high throughput of information across DCNs during such attacks. This method utilizes Forward Error Correction (FEC) coding. The original data packets are encoded by senders into more coded packets than the original amount, allowing the receiver to recover complete valid data even in the event of packet loss. Additionally, the coding module dynamically adjusts the quantity of redundant packets to minimize the occurrence of link congestion. Through NS-3 simulations, it is demonstrated that FEC coding can reduce transmission delays by approximately 35% and improve throughput by about 30% when under attack by malicious traffic, compared to other methods, as shown by the statistics.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

A Novel Adaptive Coding Technology for Data Center Network

  • Yang Rui

摘要

Due to the increasing demand for applications by users worldwide, many service providers (such as Google and Microsoft) have deployed their services on a Cross-Data Center Network (DCN) connected by a Wide Area Network (WAN). While WAN ensures interconnection between headquarters and nodes at all levels, security issues within the network are becoming increasingly prominent. In the event of a malicious attack, data transmission across DCNs may result in packet loss, leading to various adverse outcomes such as data transmission delays, low throughput, and wastage of network bandwidth. To ensure the reliability of cross-DCN traffic transmission in the face of malicious traffic attacks, this paper proposes a coding method designed to achieve low delay and high throughput of information across DCNs during such attacks. This method utilizes Forward Error Correction (FEC) coding. The original data packets are encoded by senders into more coded packets than the original amount, allowing the receiver to recover complete valid data even in the event of packet loss. Additionally, the coding module dynamically adjusts the quantity of redundant packets to minimize the occurrence of link congestion. Through NS-3 simulations, it is demonstrated that FEC coding can reduce transmission delays by approximately 35% and improve throughput by about 30% when under attack by malicious traffic, compared to other methods, as shown by the statistics.