Understand why Cisco has standardized on its three-stage design concept known as core, distribution, and access. Observe how redundancy can be utilized to increase your reliability at the edge, in aggregation points, and in the core of a network. Discover how stacking can be used to ease administration.
- [Instructor] While designing a campus network may not seem overly complex, and most often it's not, Cisco wants administrators to follow a basic design with their specific nomenclature. It's in this spirit I'll be covering what I call CDA or Core Distribution and Access. Ideally, an administrator will design a network that is predictable, redundant, documented, and easy to administer. A portion of this, as per Cisco, is accomplished by dividing the network into logical chunks, or CDA.
The access layer is easy to understand. It's where users first connect to the network. These will generally be edge switches. Often, these edge switches will provide power over Ethernet to power phones or wireless access points. Going back to easing administration, these edge switches can often be physically connected into a stack that allows them to be logically administered as if they were a single unit. A stack also supplies additional redundancies when connecting back into the network.
The access switches in the geographic area will then aggregate at the next layer, the distribution layer. Distribution switches generally are high density, high bandwidth, and have high layer three routing throughput. This is generally where VLANs aggregate that are then routed between networks. Distribution switches will have redundant links back to the core of the network. The core of the network redundantly aggregates all of the distribution switches.
These generally have very high throughput layer three routing capabilities, employ high availability techniques, and attempt to do as little filtering as possible to increase performance. The overall shape of the network really depends on the specific environment and its size. Often, some layers are collapsed. Generally, the core and distribution layers are combined into one. Redundancy can also drive a lot of decision making process when designing the network and spec'ing hardware.
If multiple access switches connect to a single distribution switch for aggregation, then this is a single point of failure. If the distribution switch goes down, all users connected to that switch are orphaned. This can be mitigated by adding redundant distribution switches. If these redundant distribution switches now connect to a single core switch, then another single point of failure has been introduced. Dual core devices can be added to mitigate this single point of failure.
Key decisions here are budget and how much availability is required for this network. Redundancy is great, but if the budget doesn't allow for it, then it's wasted time planning for it. If this is a remote network with five non-critical employees, it may be a waste of resources to have the three tier design with full redundancy. The CDA design has been Cisco's template since I began my career, and it's still alive and well in today's networks.
Having a grasp on design will serve you well in your networks and on the test.
- Switch design and basic port configuration
- VLAN operation and configuration
- VLAN creation and access ports
- VTP operation and advanced configuration
- STP operation and configuration
- EtherChannel operation and configuration
- Multilayer Switching
- SPAN and RSPAN operation and configuration
- Exploring high availability