File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: QOAG: An efficient queueing policy for input-buffered packet switches

TitleQOAG: An efficient queueing policy for input-buffered packet switches
Authors
Issue Date2000
Citation
Ieee International Conference On Communications, 2000, v. 3, p. 1753-1757 How to Cite?
AbstractAn efficient self-adaptive packet queueing policy, called Queueing with Output Address Grouping (QOAG), is proposed for optimizing the performance of an input buffered packet switch. Each input port of the N×N switch under consideration has Q queues and each queue has B packet buffers, where I
Persistent Identifierhttp://hdl.handle.net/10722/158280

 

DC FieldValueLanguage
dc.contributor.authorLiu, NHen_US
dc.contributor.authorYeung, Kwan Len_US
dc.date.accessioned2012-08-08T08:58:52Z-
dc.date.available2012-08-08T08:58:52Z-
dc.date.issued2000en_US
dc.identifier.citationIeee International Conference On Communications, 2000, v. 3, p. 1753-1757en_US
dc.identifier.urihttp://hdl.handle.net/10722/158280-
dc.description.abstractAn efficient self-adaptive packet queueing policy, called Queueing with Output Address Grouping (QOAG), is proposed for optimizing the performance of an input buffered packet switch. Each input port of the N×N switch under consideration has Q queues and each queue has B packet buffers, where I<Q<N. Using QOAG, a packet arrived at an input port is assigned to the queue which has some backlog packets with the same output address as that of the new packet. If the output address of the new packet is different from all current buffered packets in all queues, it is assigned to the shortest queue. The performance of QOAG is compared with Odd-Even queueing policy in [1] by simulations. Zipf distribution version II is used to model the non-uniform packet output distributions. We found that for a 16×16 switch with B = 20 buffers at each queue and input load p = 0.7, the mean packet delays are 58.1 and 91.2 time slots and the mean throughputs are 0.474 and 0.355 for using QOAG and Odd-Even queueing respectively. This represents a 57% cut in mean packet delay and 25% increase in throughput when QOAG is used.en_US
dc.languageengen_US
dc.relation.ispartofIEEE International Conference on Communicationsen_US
dc.titleQOAG: An efficient queueing policy for input-buffered packet switchesen_US
dc.typeConference_Paperen_US
dc.identifier.emailYeung, Kwan L:kyeung@eee.hku.hken_US
dc.identifier.authorityYeung, Kwan L=rp00204en_US
dc.description.naturelink_to_subscribed_fulltexten_US
dc.identifier.scopuseid_2-s2.0-0033685104en_US
dc.identifier.volume3en_US
dc.identifier.spage1753en_US
dc.identifier.epage1757en_US
dc.identifier.scopusauthoridLiu, NH=7402430988en_US
dc.identifier.scopusauthoridYeung, Kwan L=7202424908en_US

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats