QoS is a critical component of any IP Telephony solution. Converged networks must provide secure, predictable, measurable, and sometimes guaranteed services. In order to ensure successful end-to-end business solutions, Quality of Service (QoS) is required to manage network resources. There are several reasons why QoS is required when integrating real-time, delay-sensitive traffic such as voice and video into the network. These reasons include:
- Delay Issues
- Bandwidth Issues
- Jitter Issues
- Packet Loss Issues
All packets in a network experience some kind of delay from the time the packet is first sent to when it arrives at its intended destination. This total delay, from start to finish, is referred to as latency. There are several types of delay that may be experienced by packets or frames. Some common causes of delay include, but are not limited to serialization delay, queuing delay, processing delay, and forwarding delay.
Serialization delay refers to the amount of time that it takes to serially send bits, i.e. one bit at a time, across the wire. Queuing delay is the delay experienced when packets wait for other packets to be sent. Processing delay is the time taken by the digital signal processor (DSP) to compress a block of PCM samples. This is also referred to as Coder delay. And finally, forwarding delay includes the processing time from when a frame and when the packet has been placed in the output queue.
NOTE: You are not required to go into any detail on DSPs or PCM sampling in the TSHOOT exam. These will not be described in any further detail in the remainder of this chapter.
While there are numerous types of delay, all delay types fall into one of two categories. These two categories are fixed delay and variable delay. Fixed delay components add directly to the overall delay on the connection. Examples of fixed delay include serialization delay, processing delay, and packetization delay. Packetization delay is the time taken to fill a packet payload with encoded or compressed speech.
Variable delays, on the other hand, arise from queuing delays in the egress trunk buffers on the serial port connected to the WAN. These buffers create variable delays, called jitter, across the network. Variable delays are handled through the de-jitter buffer at the receiving router or gateway. Jitter is described in additional detail later in this section.
Generally speaking, bandwidth refers to the number of bits per second (bps) that are expected to be delivered successfully across some medium. Based on this definition, bandwidth is equal to the physical link speed or clock rate of the interface. In switching terms, however, the term bandwidth refers to the capacity of the switch fabric. Therefore, the bandwidth considerations for WAN connections, for example, are not necessarily the same for LAN connections.
Jitter is the variation in delay between consecutive packets and is caused by variable queuing delays. For this reason, jitter is often commonly referred to as variation delay. While such variations may be acceptable for applications and data traffic, they can severely impact isochronous traffic, such as digitized voice, which requires that packets are transmitted in a consistent, uniform manner. The varying arrival time of the packets can cause gaps in the recreation and playback of the original voice signal. This is both undesirable and annoying to the listener. Jitter can be mitigated using de-jitter buffers.
Packet loss occurs when one or more packets traversing the network fail to reach their intended destination. This may occur for several reasons, such as bit errors, lack of space in queues, and most commonly due to congestion. While this does not generally affect connection-oriented protocols, such as TCP, packet loss can cause major issues for real-time traffic such as voice and streaming video traffic. Packet loss can be mitigated using congestion management and / or congestion avoidance mechanisms. These are described later in this chapter.