The ABCs of QoS

Quality of Service (QoS) is a set of technologies that work on a network to guarantee its ability to dependably run high-priority applications and traffic under limited network capacity. QoS technologies accomplish this by providing differentiated handling and capacity allocation to specific flows in network traffic. This enables the network administrator to assign the order in which packets are handled, and the amount of bandwidth afforded to that application or traffic flow. 


Measurements of concern to QoS are bandwidth (throughput), latency (delay), jitter (variance in latency), and error rate. This renders QoS of particular importance to high-bandwidth, real-time traffic such as voice over IP (VoIP), video conferencing, and video-on-demand that have a high sensitivity to latency and jitter. These applications, with minimum bandwidth requirements and maximum latency limits, are called “inelastic.”

The QoS mechanisms for ordering packets and allotting bandwidth are queuing and bandwidth management respectively. Before they can be implemented however, traffic must be differentiated using classification tools. The classification of traffic according to policy allows organizations to ensure the consistency and adequate availability of resources for their most important applications.

Traffic can be classified crudely by port or IP, or using a more sophisticated approaches such as by application or user. The latter parameters allow for a more meaningful identification, and consequently, classification of the data.

Next, queuing and bandwidth management tools are assigned rules to handle traffic flows specific to the classification they received upon entering the network. 

The queuing mechanism allows for packets within traffic flows to be stored until the network is ready to process it. Priority Queuing (PQ) is developed to ensure the necessary availability and minimal latency of network performance for the most important batches of applications and traffic by providing an assigned priority and specific bandwidth to them based on their classification. This ensures the most important activities on a network are not starved of bandwidth by activities of lower priority. Applications, users, and traffic can be batched in up to 8 differentiated queues.

Bandwidth management mechanisms measure and control traffic flows on the network to avoid exceeding its capacity and the resulting network congestion that occurs. Mechanisms for bandwidth management include traffic shaping, a rate limiting technique used to optimize or guarantee performance and increase usable bandwidth where necessary, and scheduling algorithms, which offer varied methods for providing bandwidth to specific traffic flows.

Depending on the provider, the above services and controls can be managed and consolidated down to a single box. Such is the case for QoS via Palo Alto Networks firewalls. Thus, to communicate QoS measures and classification outside the box and downstream network infrastructure Differentiated Services Code Point (DSCP) is implemented. DSCP marks each packet based on its classification, and communicates this to each box the packet travels through, ensuring a consistent implementation of QoS policy.



Ignite 2017 Vancouver