A method for implementing a scalable data center network architecture

Abstract

The invention discloses a scalable data center network architecture, which uses optical switches to replace traditional electrical switches, and integrates racks (6), ToR switches (5), aggregation switches (4), The core switch (3) and the content switch (2) are grouped into different components, and then each component is placed remotely and distributed on a group of adjacent nodes in the backbone network, and the data between components passes through the backbone network The optical fiber in the data center is switched and transmitted, thereby effectively using the backbone network infrastructure to simplify the internal connection of the data center and enhance the scalability of the data center network architecture. The invention also discloses the implementation method of the data center network architecture. The grouping of racks, the determination of the number of switches, the placement of components, and the solution of service routing are all attributed to an integer linear programming and then solved using mathematical tools or based on heuristic algorithms to design architectures. The design is flexible and adaptable. .

Description

technical field [0001] The present invention relates to a data center network architecture, in particular to a method for realizing a scalable data center network architecture. Background technique [0002] A data center is a business support platform for cloud services and consists of a large number of servers. These servers are organized in the form of racks (Rack), and dozens to thousands of servers are usually placed on each rack to provide parallel computing and resource storage. There are data exchange requirements between servers on the same rack and different racks. To meet this need, the data center contains a multi-level switching network inside. [0003] Usually, a ToR (Top-of-Rack) switch is placed on the top of each rack in the data center, which is responsible for data exchange between servers in this rack and data exchange interfaces with other racks. ToR switches usually need to have packet header parsing and routing functions. Each ToR switch is connecte...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a method for implementing a scalable data center network architecture. The data network center adopts a distributed architecture, which solves the difficulty of power supply and heat dissipation caused by centralized placement in the traditional architecture. High cost and insufficient space, etc., with stronger scalability and lower energy consumption; two implementation methods are used to design and adjust the data center network architecture, with flexible design and strong adaptability

Images