CCNA RnS #16: Campus vs Data Center Network


CCNA RnS 16: Campus Network vs DC Network

In this article, I am going to talk about the two popular network designs Campus Network and the Data Center Network (DCN)

Ethernet LAN connects user devices to the switches and switches to other switches. This part of the LAN is called the campus network. Campus networks provide connections to the end-users. These switches further connect to other switches so that multiple locations can be connected.

Figure 1: Campus Network

Figure 1, above is a example of a campus network where end-users are connected to a switch that is installed in the wiring closet. Switches that are near to the end-users are called the Access switches. Depending on the size of the network, there can be multiple access switches providing connectivity to various parts of the campus network.

These access switches further connect to another switch called the Aggregation/Distribution switch. The distribution switch is a central point where all the access switches get connected. To simplify the large network design the whole campus network is designed in this way. This 3 tier design helps administrators to scale the network without disturbing the other part of the network and isolate the problem in the network.

Coming back to the campus network, aggregation/Distribution switches gets connected to the core switches. These core switches are the gateway for a campus network. Gateway means anything that needs to cross the campus network should pass through the core switch. For example, traffic from the end-user to the application server should cross to the core switch.

Enterprises build their network to provide access to the applications. These application servers are high compute and high throughput devices. Sometimes they have different protocols to make communication happen between the servers – a good example is traffic between the Database server to storage may or may not be on Ethernet. It could be over Fiber Channels sometimes known as Storage Area Networks (SAN). Therefore, the connectivity design for servers is not the same as for end-users. The network setup for application server connectivity is called the Data Center Network.

In the Data Center Network, Applications servers are connected to Top of Rack (ToR) switches these switches are installed within the RACK where servers are installed, and ToR is connected to other switches EoR (End of Row) or MoR (Middle of Row) switches. As we require high throughput, we need more processing power on switches, 10G, 40G, 100G, or more bandwidth Interfaces.

Figure 2: Data Center Network (DCN)

The Data Center network is different from the campus network in terms of throughput, environmental conditions, and the high availability requirement. DCN devices are installed in a controlled environment sometimes called the server room or data center.

As you can see in Figure 2, Servers are installed in physical racks connected to ToR switches these core switches connect to MoR or EoR switches. The MoR and EoR switches get connected to the Data Center Network Core SW (DCN SW). The requirement for MoR, EoR or ToR depends on the size of the network. Some network can be built without MoR or EoR. 

Now as already stated the whole idea to build a network is to do business where business applications are installed in Data Center and access to these business applications is provided to the end users. This is the very basic approach for an organization to adopt digitization. To allow traffic from end-users to applications these two networks are built as mentioned in Figures 1 and 2. Finally, these networks get connected where DCN SW and Campus Core SW are connected.

To forward traffic from an end-user to a server and vice versa, each switch performs some logic. The next one or two articles will cover switching traffic forwarding logic. Hope you find this informative!

  Continue Reading...