HS-DCell: A Highly Scalable DCell-Based Server-Centric Topology for Data Center Networks

被引:1
|
作者
Wang, Guijuan [1 ]
Zhang, Yazhi [1 ]
Yu, Jiguo [2 ,3 ]
Ma, Meijie [4 ]
Hu, Chunqiang [5 ]
Fan, Jianxi [6 ]
Zhang, Li [1 ]
机构
[1] Qilu Univ Technol, Shandong Acad Sci, Key Lab Comp Power Network & Informat Secur, Minist Educ,Shandong Comp Sci Ctr, Jinan 250353, Peoples R China
[2] Univ Elect Sci & Technol China, Sch Comp Sci & Engn, Chengdu 250353, Peoples R China
[3] Qilu Univ Technol, Big Data Res Inst, Jinan 250353, Peoples R China
[4] Qilu Univ Technol, Shandong Acad Sci, Sch Math & Stat, Jinan 250353, Peoples R China
[5] Chongqing Univ, Sch Big Data & Software Engn, Chongqing 400044, Peoples R China
[6] Soochow Univ, Sch Comp Sci & Technol, Suzhou 215006, Peoples R China
基金
中国国家自然科学基金;
关键词
Servers; Data centers; Scalability; Routing; Fault tolerant systems; Fault tolerance; Computer science; Data center network; server-centric; scalability; fault tolerance; routing algorithm; PORT; TECHNOLOGIES; ARCHITECTURES; PATHS; EDGE;
D O I
10.1109/TNET.2024.3398628
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Topology design is vital to the high performance data center networks. Due to the limited scalability, many traditional server-centric data center networks are confronting the updating and upgrading hurdles. To address the issue, this paper proposes a highly scalable DCell-based server-centric data center network topology, called HS-DCell, which can use inexpensive and typical switches and servers with only three network ports to achieve excellent network performance HS-DCell can accommodate a large number of servers, and its diameter increases linearly with the growth of network levels, which is better than that of most existing server-centric networks. Furthermore, a fault-free routing algorithm and a fault-tolerant routing algorithm are developed based on HS-DCell. Compared with other mainstream server-centric network topologies, the experimental results show that HS-DCell has obvious advantages in many key performance indicators including scalability, fault tolerance, and server port utilization.
引用
收藏
页码:3808 / 3823
页数:16
相关论文
共 11 条
  • [11] Energy-aware Coflow and Antenna Scheduling for Hybrid Server-Centric Data Center Networks
    Li, Tao
    Santini, Silvia
    2017 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS (ICC), 2017,