WAN Options

When wide area networks (WANs) are required, there are three broad categories the network designer has to choose from:

• Private networks

• Leased networks

• Public networks

Private networks are relatively rare. They are networks where the WAN customer actually owns the WAN infrastructure. This includes cabling, fiber, switching equipment, etc. Leased networks are the most common; they are wide area services purchased from a service provider. Public networks generally involve attaching sites to the public Internet and allowing them to communicate across the public space.

Cisco supports each of these different types of WANs, providing ample flexibility and many options for meeting internetworking business requirements. Let’s consider each of these WAN types.

Private Networks

As mentioned, private networks are relatively rare due to their cost and lack of flexibility. Few enterprises have the capacity to install truly private networks. Do you really want to know what it would cost to run fiber from Denver to Chicago? A cost-effective example of a private network is a large university (or other) campus where the distances are relatively short. In this case, the cost of installing the fiber or copper to connect separate buildings is relatively small compared to the cost of paying an external service provider to connect buildings only a few hundred yards apart. Another example of a private network might include wireless technologies. Your humble author has installed more than one microwave system capable of covering significant distances.

While the up-front costs of private networks can be considerable, there are advantages. Once installed, operating costs are relatively low compared to other offerings, and maintenance and management are simplified in that no external entity is involved.

Leased Networks

There are many technologies for Layer 2 and Layer 3 that are available in the marketplace when a leased WAN network is required. However, they can be consolidated into several large categories. This section discusses three of these categories:

• Dedicated leased lines

• Circuit-switched networks

• Packet-switched networks

Most of the available technologies fit into one of these three categories.

Dedicated Leased Lines

Point-to-point serial links are dedicated links that provide full-time connectivity. They connect into a CSU/DSU (channel service unit/data service unit), which then plugs into the demarc provided by the telephone company.

Figure 6.1 shows how point-to-point connections can be made between remote offices and the corporate office.

Figure 6.1: Point-to-point connections between branches and the corporate office

If the established business requirements dictate that constant connection and steady data flow must prevail, a dedicated point-to-point connection can be an optimal solution. But even so, this approach does have a disadvantage associated with it: Point-to-point connections require that you pay tariffs for them even when the connection is in an idle state with no data being transmitted. Essentially, you pay for the entire pipe whether you use it or not.

Dedicated lines are always the preferred choice when they are available and cost-effective. However, in many situations dedicated lines tend to be significantly more expensive than other alternatives.

Circuit-Switched Networks

The fundamental advantage of circuit-switched networks is cost savings. When you hear “circuit-switched,” think “phone call.” This type of WAN connection literally makes a “call” when data needs to be transferred and then “hangs up” when finished. Since you only pay for the time you use and not for the time you do not use, this solution can be more cost-effective than the dedicated leased lines discussed earlier.

Any user with an asynchronous modem can make connections to an internetwork using the PSTN (Public Switched Telephone Network), and there are many common situations requiring this type of access. Users who are traveling can dial in to gather e-mails or update a database, while others may want to dial in from home to finish projects, send and retrieve e-mails, and even print documents to be available for them when they arrive at work the next day.

ISDN is another common circuit-switched network. ISDN interfaces can be activated when needed and then allowed to disconnect when they’re no longer required. One common use of ISDN is as a backup when another network fails. Since it is circuit-switched, it can be a very cost-effective backup, assuming your primary connection is reasonably stable!

Packet-Switched Networks

Packet-switched networks (PSNs) are essentially shared networks. In this case, it is almost like you are getting a dedicated leased line and then “selling off” the bandwidth you don’t use. You do this by sharing the infrastructure with other customers of the service provider.

Packet-switched services are usually run over a publicly maintained network such as the PSTN, but if necessary, a large organization can build a PSN. PSN data delivery can take place within frames, packets, or cells, and occurs transparently to end users.

PSNs can be either private or public networks. Switching devices forward packets using an internal addressing scheme, which can be entirely different from what’s used on the LAN. This is because a switch that’s located at the WAN provider’s office typically checks the address field of the packets only and then carries out the forwarding based on the static routes configured by an administrator.

Frame Relay is a common packet-switched technology used in networks today. It allows your organization the ability to pay for a certain amount of bandwidth (CIR), and then use more bandwidth if it’s available (bursting). For many longer distance situations, Frame Relay can be significantly less expensive than dedicated leased lines. In addition, it can be indistinguishable as far as application traffic is concerned, depending on the application, of course.

Public Networks

Most of you are aware that the most commonly used public network is the Internet. In many cases, it is significantly cheaper to connect two geographically dispersed sites to a local ISP than to invest in any of the WAN technologies discussed thus far in this chapter. These sites can then use the public Internet to carry their data back and forth. The advantages of the Internet include reduced cost, but of course, this comes at a price. There are no guarantees for delivery across the Internet; connectivity is strictly a best-effort affair.

Common technologies used when connecting to the Internet include DSL and cable modems. Virtual private network (VPN) technology is commonly used to protect data crossing the public network.