MBDC: Low Latency and Cost-effective Data Center Network Architecture
With the rapid development of information technologies such as cloud computing, big data, artificial intelligence, and edge computing, data centers have become essential infrastructure supporting the modern information society. When constructing data center networks, as the network scale increases, both latency and cost also grow. Therefore, it is essential not only to consider network scalability but also to focus on link overhead and communication latency. The hypercube is an excellent base topology for constructing data center networks. The Möbius cube, a version of the hypercube, not only retains the hypercube’s favorable properties, but also outperforms it in terms of link overhead and network diameter. In this paper, we propose a new server-centric data center network architecture, called MBDC, which is based on the Möbius cube. For networks of the same scale, MBDC achieves a smaller diameter than most existing server-centric networks. Additionally, we present an adaptive fault-tolerant routing scheme for MBDC, which is based on an improved local security information model. Extensive evaluations demonstrate that MBDC is an attractive data center network for constructing low-latency and cost-effective data centers.
Authors
- Guijuan Wang (ORCID: https://orcid.org/0000-0002-5930-9809)
- Mengjie Lv (ORCID: https://orcid.org/0000-0002-1173-3766)
- Anming Dong (ORCID: https://orcid.org/0000-0001-7470-5159)
- Li Zhang (ORCID: https://orcid.org/0000-0002-4535-3200)
- Jiguo Yu (ORCID: https://orcid.org/0000-0001-6451-1158)
- Qian Yu (ORCID: https://orcid.org/0000-0002-2253-1449)
Institutions
- University of Electronic Science and Technology of China (CN)
- Qilu University of Technology (CN)
- Nanjing University of Posts and Telecommunications (CN)
- Shandong Academy of Sciences (CN)
Publication Details
- Published
- 2024-12-13
- DOI
- https://doi.org/10.1109/hpcc64274.2024.00214
- Citations
- 2
- Primary Topic
- Cloud Computing and Resource Management
- Type
- article
- Field-Weighted Citation Impact
- 1.56
Funders
- National Science Foundation
- Qilu University of Technology