A measurement-based admission control algorithm for integrated services packet networks

Abstract

Abstract-Many designs for integrated services networks offer a bounded delay packet delivery service to support real-time applications. To provide bounded delay service, networks must use admission control to regulate their load. Previous work on admission control mainly focused on algorithms that compute the worst case theoretical queueing delay to guarantee an absolute delay bound for all packets. In this paper we describe a measurement-based admission control algorithm for predictive service, which allows occasional delay violations. We have tested our algorithm through simulations on a wide variety of network topologies and driven with various source models, including some that exhibit longrange dependence, both in themselves and in their aggregation. Our simulation results suggest that measurement-based approach combined with the relaxed service commitment of predictive service enables us to achieve a high level of network utilization while still reliably meeting delay bound. I. BOUNDED DELAY SERVICES AND PREDICTIVE SERVICEThere have been many proposals for supporting real-time applications in packet networks by proSugih Jamin was supported in part by the Uniforum Research Award and by the Office of Naval Research Laboratory under contract N00173-94-P-1205. At USC, this research is supported by AFOSR award number F49620-93-1-0082, by the NSF small-scale infrastructure grant, award number CDA-9216321, and by equipment loan from Sun Microsystems, Inc. viding some form of bounded delay packet delivery service. When a flow requests real-time service, it must characterize its traffic so that the network can make its admission control decision. Typically, sources are described by either peak and average rates [FV90] or a filter like a token bucket [OON88]; these descriptions provide upper bounds on the traffic that can be generated by the source. The traditional real-time service provides a hard or absolute bound on the delay of every packet; in [FV90], [CSZ92], this service model is called guaranteed service. Admission control algorithms for guaranteed service use the a priori characterizations of sources to calculate the worst-case behavior of all the existing flows in addition to the incoming one. Network utilization under this model is usually acceptable when flows are smooth; when flows are bursty, however, guaranteed service inevitably results in low utilization [ZF94].Higher network utilization can be achieved by weakening the reliability of the delay bound. For instance, the probabilistic service described in [ZK94] does not provide for the worst-case scenario, instead it guarantees a bound on the rate of lost/late packets based on statistical characterization of traffic. In this approach, each flow is allotted an effective bandwidth that is larger than its average rate but less than its peak rate. In most cases the equivalent bandwidth is computed based on a statistical model [Hui88], [SS91] or on a fluid flow approximation [GAN91], [Kel91]) of traffic. 1 If one can precisely characterize traffic a prior...

Distributed Admission Control in Multiservice IP Networks: Concurrency Issues

“…In the context of Intserv, MBAC has been proposed to assist the provision of a predictive service for tolerant applications able to accommodate occasional delay-bound violations [7]. As the behavior of existing flows is determined by measurements rather than by their rate reservations (e.g., worst-case parameters), the service provided is less reliable due to traffic fluctuations.…”

“…However, this allows improved AC flexibility and takes advantage of statistical multiplexing, which may lead to significant utilization gains. In [7], AC is distributed node-bynode, using rate and/or delay-based equations. Other relevant MBAC algorithms are presented in [8].…”

Adaptive Service in Mobile Computing Environments

“…Predictive service bases its QoS assurances (reliable estimates as opposed to guarantees) on measured values as opposed to a priori worst-case characterization of flows [9,4]. In most related work, measurements have typically pertained to the flows, while the channel capacity is typically assumed to be invariant.…”

Capacity reservation for multimedia traffic

“…Much research on tra c control has focused on the provision of delay and loss guarantees in conjunction with statistical multiplexing (see, for instance, [9,5]). It may therefore seem a step backwards to propose a capacity reservation which enforces a deterministic service rate for a source.…”