Scalable Data Center Networks
Firat Atagun
Abstract
Data centers offer many benefits for hosting and colocating applications, data storage, servers and services. As the demand to data center’s services and resources increases; number of servers and hardware in the data centers increases exponentially. Interconnecting exponentially increasing number of services, hardware and servers while offering scalability, reliability, efficiency, fault tolerance, easily maintainability, and high network capacity becomes a challenge for data center networks. Other difficulties while designing and deploying a data center network are physical expansion, routing and forwarding inside the data center which should be scalable, reliable, and fault tolerant. In this paper I will research the characteristics, similarities and differences of approaching scalable data center network that are studied in the following three papers, [1] DCell: A scalable and fault tolerant network structure for data centers; recursively defined network architecture is explained to interconnect servers while scalability, fault tolerance and network resource usage is utilized. In [2] Portland: A scalable fault tolerant layer 2 data center network fabric; offers scalable, fault tolerant data center network design with conventional but improved data center network by using “Fat Tree†based structure. In [3] VL2: A scalable and flexible data center network; scalability, reliability and fault tolerance of a new approach to data center network is explained by network transparency, abstraction, load balancing and incremental expansion methods.
General Terms:
Design, Algorithms, Scalability, Fault Tolerance
Keywords:
Data center network, Scalability, Fault tolerance
Presentation : Scalable Data Center Networks
Not logged in