QoS Parameters

---

The important parameters that define QoS are bandwidth, link quality, delay, and jitter. However, not all applications have strict requirements for all of these parameters. For example, voice communication (telephony) has relaxed requirements for bandwidth (from about 8 to 64 Kbps), but has very demanding requirements for latency (from 5ms to 150 ms or so for packet communications) and jitter (5 ms).

We will now examine the four primary QoS parameters.

Bandwidth is the measurement of the communication speed at the OSI's Physical Layer. Typically it is measured in bits per second, although sometimes frames-per-second can be used. A frame, however, is an application dependent unit and varies across video, voice, and audio applications. In general, if the wireless network is to be used for the delivery of multimedia applications, it must offer guaranteed bandwidth provisions for uninterrupted delivery of the audio/video applications. Depending on the quality of video, bit rates from 300 Kbps (e.g. streaming video) to 20 Mbps (e.g. high definition video such as HDTV) are needed. This isn't as difficult as it sounds.

Link quality refers not only to the amount of bandwidth that is needed to enable quality connectivity, but also to the ability to ensure that a significant number of bits (or frames or packets) arrive error-free. The ideal network would also provide the ability for error-riden frames or packets to be corrected through such mechanisms as automatic repeat request (ARQ), frame-error-rate, or packet-error-rate. Since wireless networks are designed to be just 90 percent error-free, this can present a real quandary. Link quality is usually measured in bit-error-rate and is expressed as a negative exponent of 10 (e.g. 10-9).

Delay, also known as latency, refers to the time difference between the transmission and reception of packets of information, from the source to its destination. Interactive applications, such as telephony, online gaming, and videoconferencing, are particularly sensitive to delay. For such applications, a delay of less than 100 milliseconds (ms) is needed end-to-end. Here it is important to note that the wireless network receives only a portion of this end-to-end delay budget. Values in the range of 5 ms to 10 ms typically are allocated for the wireless network portion of the link.

Jitter, also known as delay variation, refers to the variation in arrival (latency) of data packets. Jitter is an important consideration in isochronous applications (i.e. applications whose signals are dependant on some uniform timing—many carry their own timing information embedded as part of the signal), because these applications must generate or process regular amounts of data at fixed intervals. Examples of such applications include telephones that send and receive voice samples at regular intervals, and fixed-rate video that generates data at regular intervals and thus must receive data at regular intervals. Often, the method used to minimize jitter employs buffers to control the delay variation at the receiver end. This method, however, translates the jitter problem into a latency problem, and thus is only useful for non-interactive applications such as streaming video. A typical maximum jitter requirement for high quality video is about 10 ms— this is difficult for a WLAN to provide, but not as troublesome as provisioning link quality over a WLAN environment.

---