Quality of Service

With no policies in place, network channels function on a best-effort delivery basis – that is, all the traffic across a channel would have equal priority, and so traffic packets for all applications/servers/users have the same chance of being delivered, or being dropped. In other words, since all traffic has equal priority, administrators have no flexibility in controlling traffic flow.

This is where Quality of Service (QoS) policies would help. With QoS policies in place, the administrator would be able to allocate bandwidth to business-critical applications like CRP, VoIP, databases etc, and conversely, ensuring available bandwidth is not encroached by gaming, streaming and other non-business applications. QoS policies often can be made to the following techniques:

• Policing: Allocate minimum bandwidth (ensuring an application gets at least a pre-defined amount of bandwidth – ideal for business critical applications). Conversely, impose maximum bandwidth limit (ensuring the application does not get more than a defined amount of bandwidth – especially for non-business related applications such as gaming, video streaming, download sites etc).
• Priority queuing: allocate bandwidth to prioritized applications, before doling out to others. This is especially useful in latency intolerant critical traffic, such as VoIP. The QoS policy can be made to identify voice traffic as Low Latency Queuing (LLQ), and therefore ensure it is always transmitted ahead of other traffic.