Key Point

The three functions that comprise the hierarchical network design model are as follows:

• Access layer Provides user and workgroup access to the resources of the network

• Distribution layer Implements the organization's policies, and provides connections between workgroups and between the workgroups and the core

• Core layer Provides high-speed transport between distribution-layer devices and to core resources

These three layers can also be thought of as modules; each module has specific functions and can therefore be designed using the optimal devices and features to meet the specific requirements of the module.

Figure 1-5 illustrates a simple network and shows how it maps to the hierarchical model. (Later chapters in this book detail the functions of the devices shown in this figure.)

Figure 1-5. The Hierarchical Network Design Model as Mapped to a Simple Network

Figure 1-5. The Hierarchical Network Design Model as Mapped to a Simple Network

Do you always need to have separate devices for each layer? No. Consider how the Transmission Control Protocol/Internet Protocol (TCP/IP) suite is an implementation of the OSI model. The TCP/IP model combines some of the OSI layers; for example, the TCP/IP application layer represents the OSI model application, presentation, and session layers. Similarly, your implementation of the hierarchical model can combine some of the functions into one physical device, especially if you have a smaller network.

Some factors to consider when designing each of the hierarchical layers are described in the following sections.

Access Layer

The access layer is where users access the network. Users can be local or remote.

Local users typically access the network through connections to a hub or a switch. Recall that hubs operate at OSI Layer 1, and all devices connected to a hub are in the same collision (or bandwidth) domain. Switches operate at Layer 2, and each port on a switch is its own collision domain, meaning that multiple conversations between devices connected through the switch can be happening simultaneously. Using a LAN switch rather than a hub has a performance advantage: A LAN switch forwards unicast traffic only out of the port through which the traffic's destination is considered reachable. However, a hub forwards all traffic out of all its ports. For this reason, most of today's networks have LAN switches rather than hubs. (Switching, including Layer 3 switching, is discussed in Chapter 2, "Switching Design.")

Remote users might access the network through the Internet, using VPN connections, for example. Connections to the Internet can be through dial-up, digital subscriber line (DSL), cable, and so forth. Other access possibilities include WANs such as Frame Relay, leased lines, and Integrated Services Digital Network (ISDN).

The access layer must also ensure that only users who are authorized to access the network are admitted.

Distribution Layer

The distribution layer interfaces between the core and access layers, and between access layer workgroups.

The distribution layer functions and characteristics include the following:

• Implementing policies by filtering, and prioritizing and queuing traffic.

• Routing between the access and core layers. If different routing protocols are implemented at these other two layers, the distribution layer is responsible for redistributing (sharing) among the routing protocols, and filtering if necessary (as discussed in Chapter 3, "IPv4 Routing Design").

• Performing route summarization (as also discussed in Chapter 3). When routes are summarized, routers have only summary routes in their routing tables, instead of unnecessary detailed routes. This results in smaller routing tables, which reduces the router memory required. Routing updates are also smaller and therefore use less bandwidth on the network. As discussed in Chapter 3, route summarization is only possible if the IP addressing scheme is designed properly.

• Providing redundant connections, both to access devices and to core devices.

• Aggregating multiple lower-speed access connections into higher-speed core connections and converting between different media types (for example, between Ethernet and Frame Relay connections), if necessary.

Core Layer

The core layer provides a high-speed backbone. Functions and attributes of the core layer include the following:

• Providing high-speed, low-latency links and devices for quick transport of data across the backbone.

• Providing a highly reliable and available backbone. This is accomplished by implementing redundancy in both devices and links so that no single points of failure exist.

• Adapting to network changes quickly by implementing a quick-converging routing protocol. The routing protocol can also be configured to load-balance over redundant links so that the extra capacity can be used when no failures exist.

Filtering is not performed at this layer, because it would slow processing. Filtering is done at the distribution layer.

Limitations of the Hierarchical Model

The hierarchical model is useful for smaller networks, but it does not scale well to larger, more complex networks. With only three layers, the model does not allow the modularity required to efficiently design networks with many devices and features. The Enterprise Composite Network Model, introduced in the following section, provides additional modularity and functions.

Was this article helpful?

0 0

Post a comment