wilamowski-b-m-irwin-j-d-industrial-communication-systems-2011

Quality of Service 19-5
segment data (if needed), encode signals (if necessary), process data, and turn data into packets;
serialization delay, the time that takes to place the bits of a packet, encapsulated in a frame, onto
the physical media; the propagation delay or time that it takes to transmit the bits of a frame
across the physical wire; the processing delay, the time that takes a network device to take the
frame from the input interface, place it into a receive queue, and place it into the output queue of
the outgoing interface; and queuing delay, the amount of time that a packet resides in the output
queue of an interface.
• Delay variation: Also named “jitter” is the difference in the end-to-end delay between packets. If,
for example, one packet requires 100.ms to traverse the network from the source endpoint to the
destination endpoint, and the following packet requires 130.ms to make the same path, the delay
variation between packets is calculated as 30.ms. Many devices in a network have a jitter buffer for
smoothing out changes in arrival times of data packets containing, for example, voice or video.
• Packet loss: It is a measure of packets transmitted and received compared to the total number that
were transmitted and expressed as the percentage of packets that were dropped. Tail drops occur
when the output queue is full, being not the unique drops but the most frequent ones, and the
origin of the drops is the congestion in the link.
It is important to see that many applications can deliver unpredictable bursts of traffic. It is practically
impossible, for example, to predict the usage patterns for Web or file transfer, but network professionals
need to be able to support critical applications during peak periods.
QoS technology permits to do the following:
• Predict response times for end-to-end network services
• Manage loss in times of inevitable bursty congestion
• Manage jitter or delay sensitive applications
• Set traffic priorities across the network
• Support dedicated bandwidth
• Avoid or manage network congestion
To achieve these results, it is necessary to do a constant monitoring, real-time monitoring, a careful
traffic engineering and, once we get these features, it is possible to create a good planning for every
application.
Each of them will require a defined service level. The objectives can be resumed:
• All network traffic must reach its service levels. For example, keeping the latency below a certain
value for the voice over IP traffic to get a good quality.
• During the periods of network congestion, the most important traffic must always enjoy the
resources it needs.
• As an added benefit, these techniques allow to optimize the use of network resources, delaying the
need to spend more money to add resources.
There are a number of processes [2,12] involved in QoS:
• Classifying the traffic: A descriptor value must be used to categorize a packet or frame within a
specified group. This will make the packet accessible for QoS handling in the network or within a
network device. The classification process will allow segmenting the traffic into multiple priority
levels or classes of service. Another possible approach is “QoS signaling” in which the classification
is done based on application-flow rather than a packet basis.
• Traffic shaping: It is used to create a traffic flow for limiting the bandwidth potential of the flow. It
determines if a packet is in or out of profile by comparing the traffic rate to the configured police,
which limits the bandwidth consumed by a flow of traffic. The traffic exceeding the configured
rate is buffered first in an attempt to minimize loss.
• Marking is the process of setting or changing a priority value of a frame or packet.
© 2011 by Taylor and Francis Group, LLC