Wide Area Network (WAN)

As the network expands to multiple locations, there is a need to connect the multiple campus or branch networks to each other. This connectivity from the campus networks from different locations to the data center might be on dedicated links taken from service providers or over a shared infrastructure of the service provider, who will provide virtual network services over the Service Provider (SP) infrastructure and isolate traffic of different customers. An alternative way of connecting the branches to the data center could be over the internet over manual tunnels configured using encryption.

It might seem that the role of the wide area network is just to transport packets between two locations, which is largely true. The complexity starts to set in when we start factoring in the fact that the WAN bandwidths are much less when compared to the bandwidths available in the campus. Consider the fact that campus or the LAN might have 10G or even 40G backbones, but the WAN links would generally be an order of magnitude lower, primarily due to the cost of the WAN links. This forces the network administrators to ensure that these links are optimally utilized, while providing services to the end users with an acceptable quality. This translates to segregating the traffic and prioritizing the critical traffic from nonbusiness traffic and ensuring that the business-critical traffic takes priority over the WAN links.

Redundancy is an essential element of any network design as discussed in the previous chapters. WAN links are more likely to fail, compared to the links within the campus. Also, any outages on the WAN links generally take more time to be restored. Any outage in a WAN link between two locations would isolate the two locations, and, if the business applications cannot work in a standalone mode, the business would be impacted for the duration of time the WAN links are down. To ensure a high uptime of the network and services, network designers use redundant links or use multiple service providers to provide the WAN connectivity. Some networks might also use the internet as the fallback option. These options are depicted in the following figure:

Going overboard with redundancy can become a problem of plenty, as the network designers have to ensure that controls are in place for the network performance and traffic flows to be deterministic. This means that the network design has to be done in a manner such that certain traffic always takes a certain path or link/links as the first preference and only when the primary path/link is down, should the traffic switch to the backup link. This is achieved by defining and configuring the right routing protocols and policies on the WAN links.

The main functions of the WAN block are summarized as follows:

• Providing connectivity between networks at different physical locations
• Providing differentiated services and the right quality of services to the different types of traffic
• Providing enough redundancy and ensuring that the network flows stay deterministic in case of link failures
• Exchanging routes between the locations, or with the SP in case of virtual private networks to direct certain types of traffic on certain links

We will cover these functions and the methods of deploying the different types of WAN connectivity in Chapter 6, Understanding and Configuring WAN Technologies.