Quality of Service on the Internet: Fact, Fiction, or Compromise?

The Internet has historically offered a single level of service, that of "best effort," where all data packets are treated with equity in the network. However, we are finding that the Internet itself does not offer a single level of service quality, and some areas of the network exhibit high levels of congestion and consequently poor quality, while other areas display consistent levels of high quality service. Customers are now voicing a requirement to define a consistent service quality they wish to be provided, and network service providers are seeking ways in which to implement such a requirement. This effort is happening within the umbrella called "Quality of Service" (QoS). Of course, this is now a phrase which has become overly used, often in vague, nondefinitive references. QoS discussions currently embrace abstract concepts, varying ideologies, and moreover, lack a unified definition of what QoS actually is and how it might be implemented. Subsequently, expectations regarding QoS have not been appropriately managed within the Internet community at large of how QoS technologies might be realistically deployed on a global scale. A more important question is whether ubiquitous end-to-end QoS is even realistic in the Internet, given the fact that the decentralized nature of the Internet does not lend itself to homogenous mechanisms to differentiate traffic. This paper examines the various methods of delivering QoS in the Internet, and attempts to provide an objective overview on whether QoS in the Internet is fact, fiction, or a matter of compromise. file:///D|/docs/gih/papers/inet98/index.html (1 of 31)22/08/2003 8:04:44 AM Quality of Service on the Internet: Fact, Fiction, or Compromise?