Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies
DOI: 10.1109/infcom.2002.1019366
|View full text |Cite
|
Sign up to set email alerts
|

Towards simple, high-performance schedulers for high-aggregate bandwidth switches

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

1
54
0
2

Publication Types

Select...
4
2
2

Relationship

0
8

Authors

Journals

citations
Cited by 74 publications
(62 citation statements)
references
References 27 publications
1
54
0
2
Order By: Relevance
“…The complexity of this algorithm adopted to module-level matching is reduced to N 1.5 , and is still high for large-scale switches. The APSARA algorithm [17] tries to speed up the scheduling process by searching in parallel for the highest weight from reduced space matchings. For example, by swapping any two connections from the matching used in the previous time slot, the APSARA algorithm can find a new matching.…”
Section: Packet Dispatching Scheme For the Mcnsmentioning
confidence: 99%
“…The complexity of this algorithm adopted to module-level matching is reduced to N 1.5 , and is still high for large-scale switches. The APSARA algorithm [17] tries to speed up the scheduling process by searching in parallel for the highest weight from reduced space matchings. For example, by swapping any two connections from the matching used in the previous time slot, the APSARA algorithm can find a new matching.…”
Section: Packet Dispatching Scheme For the Mcnsmentioning
confidence: 99%
“…From both (10) and (11) Consider the policy D and fix θ 0 = 3/2. We know, from Lemma 1, that D is F -work-conserving (in this case, γ = 0 and θ = θ 0 ).…”
Section: Property 2 It Can Never Happen That It Exists J Such Thatmentioning
confidence: 99%
“…But they do not guarantee delay bounds for each packet. Many practical scheduling algorithms [9,10] have been proposed to approximate MWM performance. Their simplicity usually leads to some performance penalties, usually in the form of throughput degradation and/or larger delays.…”
Section: Introductionmentioning
confidence: 99%
“…But the delay performance is unacceptably high. Several algorithms were proposed by Giaccone et al [1] to improve the delay performance. SERENA, which also considers new arrivals, besides using memory, has a much better delay performance than the one in [7].…”
Section: Introductionmentioning
confidence: 99%