Enterprise Core Network Design (IPv6)

The core layer is the simplest yet the most critical layer. This layer is the backbone of the network. The core needs to be highly reliable and switch high traffic loads as fast as possible. It provides a limited set of services and is highly available using redundant devices and configurations to ensure that software upgrades or hardware changes can be made without disrupting the applications. The core provides a Layer 3 routing module for all traffic in and out of the enterprise network. Routing is critical for the data center core and would need to be configured using built-in robust security mechanism to avoid incorrect neighbor peering, injection of incorrect routes, and routing loops. To prevent this problem, the core layer must include the following:

■ Route peer authentication

■ Route filtering

■ Log neighbor changes

■ Antispoofing: Unicast RPF (uRPF) and rate limiting

Table 2-1 shows some of the do’s and don’ts that can be used while designing the core of the network.

Table 2-1 Do’s and Don’ts of Core Layer Designs

Do

Don’t

Design the core for high reliability. Consider using 10 GigE and Gigabit Ethernet technologies in port-channel configurations to facilitate high speeds and redundancy.

Use software-based features because they can potentially slow down traffic.


Design core layer devices with little latency.

Support workgroup access in the core layer.

Use routing protocols with low convergence

Consider hardware/software upgrades to

time.

improve performance when adding new core nodes.

Next post:

Previous post: