High availability and contingency schemes
Networks are critical for business purposes. Lack of service in some companies, even for a few seconds, might generate a significant financial loss, or prevent the detection of real time events. Therefore, the network as a system resource must be designed in some cases with contingency features in mind.

A highly available network is not a concept that applies to hardware redundancy only. Design must also include how software is going to be configured to proceed when problems arise. A poorly designed contingency scheme may lead to lack of availability because systems might simply not converge to normal operation within the expected timeframe.

Quality of Service
Networks are the highways and crossroads for the communication of business applications. Flows of different nature –data, voice and video- converge on a common media and compete for the available resources. As a consequence, time sensitive applications like voice need to be treated according to their requirements. Lack of proper delay, jitter bounds or sequentiality could compromise the quality of a critical subset of the applications running on top the network. Network administrators can meet this challenge by applying Quality of Service (QoS) to the infrastructure.

QoS is a complex discipline. Some manufacturers offer a strong set of QoS features in their product portfolio. Traffic Shaping, Resource Reservation, Early Detection, Fair Queuing, Absolute Priority, and many other techniques are some of the tools administrators can use in order to define a QoS policy for their environments.

The applicability of some or all of these techniques is sometimes strongly dependent on the type of flow. For instance, multimedia applications need to be administered quite differently to the way we have to deal with bulk transfer applications. The challenge is how to design the QoS policy in order to guarantee that every application will have a chance of a fair amount of the available resources. This is basically what QoS is all about: the need to identify and classify traffic, and assign the resources according to enterprise policies.