QoS Technical Paper
by Galen Collins, Associate Professor at NAU
Quality of Service (QoS) is a term that refers to guaranteeing bandwidth across an entire network and the delivery of traffic without any packet loss and jitter. More specifically, it enables network managers to either reserve or prioritize traffic through an enterprise network. Merriman (1998) feels QoS will be deployed in distributed environments when the cost and added complexity of using this capability is a better alternative than increasing bandwidth.
While any multimedia object can be transmitted through a network, insufficient bandwidth will affect the quality of the transmission. This is particularly true of time-sensitive continuous media such as video. Several frames, for example, may be lost due to transient overloading and a lack of QoS control in either the network architecture or operating system.(Nakajima & Tezuka, 1994). With QoS control, the desired QoS level is specified for each application where the QoS level can be changed (upgraded or degraded) based on user requests and the number of applications and their respective importance when the system is overloaded. QoS specifications include such things as audio frequency range, audio levels, audio free interval, video frame rate, video resolutions, video error free interval, and minimum data rate for user data.
Maintaining the QoS level becomes more difficult when running real-time applications over packet-switched networks like the Internet. According to Schwantag (1997), parts of the Internet are often heavily loaded causing real-time data to deliver poor quality during periods of congestion. For example, Mbone audio multicasts are unintelligible during certain parts of the day because of more than 30% packet loss. Consequently, various protocols are being developed to support specified qualities of service (e.g., RSVP). These protocols address the problem that different applications require different services (e.g., data transfer rate requirement for a file transfer application is much less than for multimedia conferencing). The Resource Reservation Setup Protocol (RSVP), for example, reserves resources along a route where RSVP-enabled routers schedule and prioritize packets to fulfill the QoS (PC Webopedia, 1998). Another type of protocol is Differentiated Services (Diff-Serv), which uses a common language to mark or label packets with their proper priority or class to trigger QoS along the traffic's route. However, there is no guarantee that QoS will work across different enterprise IP WANs unless Internet Service Providers synchronize on QoS standards like Diff-Serv for their routers (Higgins, 1998). While protocols like RSVP and Diff-Serv could be used in a LAN, the wealth of local bandwidth today makes this unnecessary.
Most people feel that QoS will be needed in the foreseeable future until WAN bandwidth becomes more plentiful, which someday could be a reality with technological advancements.
References
Higgins, K (1998). QoS takes shape. Internetweek. September 28, 1998, Issue:734.
Merriman, D (1998) IP QoS to the desktop: Applications and key technologies. Idea Bytes. Retrieved October 1, 1998 from the World Wide Web:http:// www.gigaweb.com
Nakajimia, T. & Tezuka, H. (1994). A continuous media application supporting dynamic QoS control on real-time mach. ACM Multimedia 1994.
Schwantag, U. (1997). An analysis of the applicability of RSVP. Retrieved September 28, 1998 from the World Wide Web:http://network-services.uoregon.edu/~ursula/thesis/thesis.html
Szuprowicz, D. (1995). Multimedia networking. New York, NY: McGraw-Hill.