Abstract
Today, Internet technologies have pervaded every corner of our society. More and more people are benefiting from the Internet in one way or the other. One of the current Internet technologies that may benefit us greatly is voice over Internet protocol (VoIP). According to Hardy (2003, p. 2), VoIP is “the interactive voice exchange capability carried over packet-switched transport employing the Internet protocol.” With VoIP technology, one can call anyone in this world at a lower cost, compared to traditional telephone systems. However, VoIP technology has one significant drawback. It has a low degree of reliability. From experimental results it is known that VoIP can achieve only 98% reliability. The service down time per year for VoIP is almost 20 working days (175 hours). For most companies and government organizations, such a degree of reliability is unacceptable since the traditional telephone system can achieve 99.999% reliability with a service down time of only five minutes per year (Kos, Klepec, & Tomaxic, 2005). As a result, quality of service (QoS) is an important concept for VoIP. Using QoS, VoIP may be able to overcome its limitation in reliability. QoS is often defined as the capability to provide resource assurance and service differentiation in a network. The definition includes two important terms—resource assurance and service differentiation. Resource assurance provides a guarantee about the amount of network resources requested by the user. On the other hand, service differentiation provides higher priority of getting network resources to those applications that have critical latency constraints. Given the importance of low latency for voice communication, it is not difficult to predict that QoS will assume greater importance in the VoIP industry as this technology gains popularity in the mass market. It is reported that VoIP is aggressively growing, and this growth is expected to continue in the coming years.
TopBackground
Today’s Internet provides best-effort services to all its applications and cannot provide any resources guarantee to applications (Kurose & Ross, 2004). Let us discuss an example to illustrate this concept. Imagine that a network uses links with capacity of 2 Mbps and supports two users—John and Peter. John uses a 1.5 Mbps VoIP application and communicates with Peter. Normally, the VoIP application works because the capacity of the link is 2 Mbps (>1.5Mbps). However, when John uses the FTP application at the same time (assuming that the FTP application needs more than 0.5 Mbps), the VoIP application cannot get the amount of resources it needs. This leads to congestion in the network and time delay in voice communication. Therefore, end-to-end QoS is required for providing resource guarantee and service differentiation in order to enhance the reliability of the VoIP system (Fineberg, 2005).
Approaches to provide QoS can be divided into three categories. They are (1) at sender side, (2) inside network, and (3) at receiver side (Wang, 2001). At the sender side, one popular way is by using adaptive multi-rate speech codec. In the network, four core technologies for providing QoS are integrated services (IntServ), differentiated services (DiffServ), multiprotocol label switching (MPLS), and traffic engineering. At the receiver side, the common way for providing QoS is optimization of the design of receiver buffer. The various methods for implementation of QoS for VoIP are described in the next section.
Key Terms in this Chapter
Quality of Service (QoS): QoS is the capability of providing resource assurance and service differentiation for different applications that use network resources.
Traffic Engineering: A repetitive process of network planning and network optimization. Using this, the resources inside the network can be used more efficiently.
Differentiated Services (DiffServ): An architecture for QoS implementation. DiffServ classifies the user traffic into different classes with different priorities.
Type of Service (TOS): A 3-bit field in the current IP header of a packet. The 3 bits represent delay, throughput, and reliability. By configuring the TOS field, the treatment of a packet inside the network can be decided.
Integrated Services (IntServ): An architecture for QoS implementation. For IntServ a resource reservation is required before an application starts to work.
802.1p: A technology for providing QoS at the Ethernet level. With 802.1p, packets with lower priority level cannot be sent out if there are packets with higher priority waiting in the queue.
Resource Assurance: Resource assurance enables an application to get the amount of network resources that it requests.
Multiprotocol Label Switching (MPLS): A management tool that helps service providers to optimize traffic inside the network. It uses label switching for packet forwarding.
Service Differentiation: Service differentiation is classifying different applications. It determines the priorities of requesting network resources for different applications.