[Proceedings] IEEE INFOCOM '92: The Conference on Computer Communications 1992
DOI: 10.1109/infcom.1992.263457
|View full text |Cite
|
Sign up to set email alerts
|

Call preemption in communication networks

Abstract: The technique of preempting ongoing calls in order to accommodate new calls of greater "value" has been used in many existing networks. In this paper we investigate some problems that relate to making the best decision on which (if any) call to preempt. We provide some results on the computational complexity of these problems and on how well simple heuristic procedures can approximate the optimal strategy.

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
65
0
1

Year Published

1997
1997
2006
2006

Publication Types

Select...
6
2
1

Relationship

0
9

Authors

Journals

citations
Cited by 96 publications
(66 citation statements)
references
References 6 publications
0
65
0
1
Order By: Relevance
“…The competitive ratio of this algorithm is O(log n), where n is the number of nodes in the network. The first competitive algorithm in the throughput model was given in [10] for the case of a single-link network and extended in [9] for a line network. A competitive algorithm for general topology networks in the throughput model was presented in [4].…”
Section: Unicast Routingmentioning
confidence: 99%
“…The competitive ratio of this algorithm is O(log n), where n is the number of nodes in the network. The first competitive algorithm in the throughput model was given in [10] for the case of a single-link network and extended in [9] for a line network. A competitive algorithm for general topology networks in the throughput model was presented in [4].…”
Section: Unicast Routingmentioning
confidence: 99%
“…Authors in [8] also demonstrated that, for the connection admission problems, if the holding time of the request is unknown when the request arrives, it is impossible to find a c-competitive algorithm if throughput is considered as the performance function. In this paper, we assume that the holding time of the request (the duration of each request) becomes known when the request arrives at the system.…”
Section: Online Problem and Competitivenessmentioning
confidence: 99%
“…Since the connection requests arrive one by one over time, it manifests the typical on-line feature [6,7], i.e., the system has to make the acceptance/rejection decision at the arrival time of the request without the advanced knowledge of the future requests. To achieve higher bandwidth utilization and/or lower blocking rate of the system, the preemptive feature [8,9,10] is deployed as to provide more "valuable" connections at the expense of less profitable ones already in services, when the total bandwidth requested exceeds the system capacity so that not all requests can be served at the same time. When and how to reject/preempt connections, and simultaneously guarantee the throughput performance, are more challenging in the on-line problem domain as there is no contribution to the throughput performance if the request is rejected/preempted.…”
Section: Introductionmentioning
confidence: 99%
“…The goal of the online algorithm is to choose a path that will be used for transmitting each file, and to decide on the transmission rate. The main difference between this model and the (well-studied) models for online routing and admission control [1,3,15,16] is that here we do not assume that the sources have prespecified transmission rate requirements, i.e., we can deal with nonstreaming types of information. We will study the idealized case where the transmission delays are all zero, and data cannot be buffered along its route.…”
Section: Introductionmentioning
confidence: 99%