The Cisco Three-Layer Hierarchical Model

Most of us were exposed to hierarchy early in life. Anyone with older siblings learned what it was like to be at the bottom of the hierarchy. Regardless of where you first discovered hierarchy, today most of us experience it in many aspects of our lives. It is hierarchy that helps us understand where things belong, how things fit together, and what functions go where. It brings order and understand ability to otherwise complex models. If you want a pay raise, for instance, hierarchy dictates that you ask your boss, not your subordinate. That is the person whose role it is to grant (or deny) your request. So basically, understanding hierarchy helps us discern where we should go to get what we need. Hierarchy has many of the same benefits in network design that it does in other areas of life. When used properly, it makes networks more predictable. It helps us define which areas should perform certain functions. Likewise, you can use tools such as access lists at certain levels in hierarchical networks and avoid them at others. Let’s face it: Large networks can be extremely complicated, with multiple protocols, detailed configurations, and diverse technologies. Hierarchy helps us summarize a complex collection of details into an understandable model. Then, as specific configurations are needed, the model dictates the appropriate manner in which to apply them.
The Cisco hierarchical model can help you design, implement, and maintain a scalable, reliable, cost-effective hierarchical internetwork. 

The Cisco hierarchical model 

The following are the three layers and their typical functions: The core layer: backbone The distribution layer: routing The access layer: switching Each layer has specific responsibilities. Remember, however, that the three layers are logical and are not necessarily physical devices. Consider the OSI model, another logical hierarchy. The seven layers describe functions but not necessarily protocols, right? Sometimes a protocol maps to more than one layer of the OSI model, and sometimes multiple protocols communicate within a single layer. In the same way, when we build physical implementations of hierarchical networks, we may have many devices in a single layer, or we might have a single device performing functions at two layers. The definition of the layers is logical, not physical. Now, let’s take a closer look at each of the layers.

The Core Layer

The core layer is literally the core of the network. At the top of the hierarchy, the core layer is responsible for transporting large amounts of traffic both reliably and quickly. The only purpose of the network’s core layer is to switch traffic as fast as possible. The traffic transported across the core is common to a majority of users. However, remember that user data is processed at the distribution layer, which forwards the requests to the core if needed. If there is a failure in the core, every single user can be affected. Therefore, fault tolerance at this layer is an issue. The core is likely to see large volumes of traffic, so speed and latency are driving concerns here. Given the function of the core, we can now consider some design specifics. Let’s start with some things we don’t want to do:

• Don’t do anything to slow down traffic. This includes using access lists, routing between virtual local area networks (VLANs), and implementing packet filtering.
• Don’t support workgroup access here.
• Avoid expanding the core (i.e., adding routers) when the internetwork grows. If performance becomes an issue in the core, give preference to upgrades over expansion.

Now, there are a few things that we want to do as we design the core: 

• Design the core for high reliability. Consider data-link technologies that facilitate both speed and redundancy, such as FDDI, Fast Ethernet (with redundant links), or even ATM.
• Design with speed in mind. The core should have very little latency.
• Select routing protocols with lower convergence times. Fast and redundant data-link connectivity is no help if your routing tables are shot!

The Distribution Layer

The distribution layer is sometimes referred to as the workgroup layer and is the communication point between the access layer and the core. The primary functions of the distribution layer are to provide routing, filtering, and WAN access and to determine how packets can access the core, if needed. The distribution layer must determine the fastest way that network service requests are handled—for example, how a file request is forwarded to a server. After the distribution layer determines the best path, it forwards the request to the core layer if necessary. The core layer then quickly transports the request to the correct service. The distribution layer is the place to implement policies for the network. Here you can exercise considerable flexibility in defining network operation. There are several actions that generally should be done at the distribution layer:

• Routing
• Implementing tools (such as access lists), packet filtering, and queuing
• Implementing security and network policies, including address translation and firewalls
• Redistributing between routing protocols, including static routing
• Routing between VLANs and other workgroup support functions
• Defining broadcast and multicast domains

 Things to avoid at the distribution layer are limited to those functions that exclusively belong to one of the other layers.

The Access Layer

The access layer controls user and workgroup access to internetwork resources. The access layer is sometimes referred to as the desktop layer. The network resources most users need will be available locally. The distribution layer handles any traffic for remote services. The following are some of the functions to be included at the access layer:

• Continued (from distribution layer) use of access control and policies 
•  Creation of separate collision domains (segmentation) 
•  Workgroup connectivity into the distribution layer

Technologies such as DDR and Ethernet switching are frequently seen in the access layer. Static routing (instead of dynamic routing protocols) is seen here as well. As already noted, three separate levels does not imply three separate routers. There could be fewer, or there could be more. Remember, this is a layered approach.