Dynamic QoS Management

Quality of Service (QoS) is the ability of a network element (e.g. an application, host or router) to have some level of assurance that its traffic and service requirements can be satisfied. To enable QoS requires the cooperation of all network layers from top-to-bottom, as well as every network element from end-to-end. Any QoS assurances are only as good as the weakest link in the chain between sender and receiver. QoS does not create bandwidth. It isn't possible for the network to give what it doesn't have, so bandwidth availability is a starting point. QoS only manages bandwidth according to application demands and network management settings, and in that regard it cannot provide certainty if it involves sharing.
To provide QoS is necessary for a good management of resources: Static QoS Management and Dynamic QoS Management. Dynamic QoS Management (DQM) includes mechanisms for maintenance, control and management of the network QoS.

- QoS control mechanisms include flow scheduling, flow shaping, QoS filtering;
- QoS maintenance mechanisms include flow monitoring and flow maintenance;
- QoS management represents an adaptation of QoS regarding policy and admission control.
 For Internet, the goal of QoS is to provide some adequate services with level of predictability and control beyond the current IP "best-effort" service. Hence, QoS with a [guaranteed] service level requires resource allocation to individual data streams. The bandwidth allocated to an application in a resource reservation is no longer available for use by best-effort applications. Considering that bandwidth is a finite resource, a priority for QoS designers has been to ensure that best-effort traffic is not starved after reservations are made. QoS-enabled (high-priority) applications must not disable the mundane (low-priority) Internet applications. The worst case should be that low-priority applications simply have a lesser (slower) service, but still function. There are essentially two types of QoS available:
  • Resource reservation (integrated services): network resources are apportioned according to an application's QoS request, and subject to bandwidth management policy. RSVP provides the mechanisms to do this [IntServ].
  • Prioritization (differentiated services): network traffic is classified and apportioned network resources according to bandwidth management policy criteria. To enable QoS, classifications give preferential treatment to applications identified as having more demanding requirements. [DiffServ] provides this service. These QoS protocols and algorithms are not competitive or mutually exclusive, but on the contrary, they are complementary. As a result, they are designed for use in combination to accommodate the varying operational requirements in different network contexts.
    Because, one user gets a better service than another, QoS requires policy enforcement that will require a policy management infrastructure. Since QoS provides added value, implies a need for accounting and billing infrastructure. It is not possible to enforce policy unless the identities of network users is not established. It should assign a level of trust. This implies a need for an authentication infrastructure.
    Applications, network topology and policy dictate which type of QoS is most appropriate for individual flows or aggregates. To accommodate the need for these different types of QoS, there are a number of different QoS protocols and algorithms.

    Last modified 1 August 2000