PDF of this tutorialThe need for QoS functionality in the access portion of data networks is stronger than ever. The rising volume of data traffic, which some analysts estimate doubles every six months, guarantees that access networks will face congestion, regardless of their speed. New applications, such as packetized voice and video, are increasing the amount of delay-sensitive traffic on these networks. QoS prioritization is needed to assure that this delay-sensitive traffic traverses congested networks in a timely fashion.
Even without new real-time media applications, QoS still contributes to the DSL service provider’s product portfolio by allowing a better differentiation between data services. This differentiation allows for a suite of differently priced services that can better serve the needs of individual customers and maximize the provider’s revenue yield. Such differentiation is only possible with network technologies that incorporate QoS mechanisms, such as ATM.
ATM technology was architected from the ground up to provide standards-based QoS mechanisms. Today, ATM remains far ahead of other protocols in enabling multimedia applications and differentiated service levels. The connectionless nature of Internet protocol (IP) has hampered the development of QoS in IP networks. However, new Internet Engineering Task Force (IETF) work currently in progress may lead to standards-based IP QoS in the future. The addition of QoS features to IP–based networks through new IETF standards such as multiprotocol label switching (MPLS) will complement ATM–based local access QoS and bring end-to-end Internet and virtual private network (VPN) QoS closer to reality. These new standards make it clear that the data-networking industry is moving toward next-generation networks with QoS functionality.
QoS Enables New DSL Services
QoS functionality allows for the prioritization of some types of traffic over others when a network becomes congested. In addition, QoS also provides mechanisms for the network to control the delay characteristics (both latency and jitter) of high-priority traffic. Together, prioritization and delay control enable the transmission of real-time voice and video streams. In the past, high-quality voice could only be sent over time-division multiplex (TDM)–based networks (i.e., the public switched telephone network [PSTN]). Today, QoS in DSL–powered local loops has enabled a new service: mixed voice telephony and data service over a single copper pair.
QoS Maximizes DSL Service Providers’ Revenue Yield
In addition to enabling new services, QoS allows for differentiation and, hence, increased revenue yield from basic data services. Without QoS, carriers can offer data service at different speeds, but these nominal speeds are often not meaningful because they represent the maximum possible transmission rate, not a guaranteed minimum throughput. With QoS, carriers can guarantee specific data rates to their clients, thus more finely differentiating between services.
Consider the examples in the figure below. Each graph shows the basic price versus quantity demand curve. At higher prices, fewer subscribers choose to purchase the service, while at lower prices the number of customers increases. The area under the curve represents a 100-percent revenue yield from the market: customers are paying the maximum they are willing to pay. For DSL data service, higher prices translate into higher guaranteed data rates (high QoS), and lower prices represent lower guaranteed rates (see Figure 1).
Figure 1. Differentiated Services Yield Highest Revenues
Revenues
Without QoS mechanisms in a network, a carrier has two basic choices: over-engineer capacity on the network so that all customers get a high level of service, or under-engineer so that all customers receive poor service on a congested network. Providing a high level of service at a high price is depicted in Figure 1 on the left. Only a few subscribers will choose to pay the high price of this premium service, leaving a large amount of potential revenue unrealized. Providing a low level of service at a low price is shown in the center. This mass-market approach produces a large customer base but not a good revenue yield.
With QoS, a carrier can offer different levels of service to customers at different prices, all on the same network. By setting each client’s QoS parameters based on his or her monthly fee, the carrier can provide as much guaranteed bandwidth to each customer as is desired. Likewise, the carrier is receiving from each customer the maximum service fee the customer is willing to pay, thus maximizing the overall revenue yield from the market. The right section of Figure 1 illustrates the increased revenue yield obtainable with four distinct service offerings; note that the yield is significantly higher than with either single-offering approach. This case is analogous to having a car dealership that sells everything from a moped to a Mercedes. For whatever amount the customer wishes to pay, the carrier has a service to offer. The carrier can cater to the customer’s specific needs, and no customers will be forced to go to a competing carrier to find that for which they are searching.
