Improving a Fault-Tolerant Routing Algorithm Using Detailed Traffic Analysis

Nayebi, Abbas; Shamaei, Arash; Sarbazi-Azad, Hamid

doi:10.1007/978-3-540-75444-2_71

Cited by 1 publication

(1 citation statement)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These metrics first introduced by authors in [23]. Some of the metrics that can be considered are: AUC(Average Utilization of all incoming and outgoing physical Channels of a node), ANB(Average Number of Blocked messages in routing element), AWB (Average Waiting time in Blocking queue), AVC(Average number of used Virtual Channels in the node switch).…”

Section: Per-node Traffic Metricmentioning

confidence: 99%

Fault-tolerance improvement of planar adaptive routing based on detailed traffic analysis

Shamaei¹,

Nayebi

Sarbazi-Azad

2007

2007 22nd International Symposium on Computer and Information Sciences

View full text Add to dashboard Cite

Currently, some coarse measures like global network latency are used to compare routing protocols. These measures do not provide enough insight of traffic distribution among network nodes in the presence of different fault regions. This paper presents a detailed traffic analysis of fault-tolerant planar adaptive routing (FTPAR) algorithm achieved by an especially developed tool. Per-node traffic analysis illustrates the traffic hotspots caused by fault regions and provides a great assistance in developing fault tolerant routing algorithms. Based on such detailed information, a simple yet effective improvement of FTPAR is suggested. Moreover, the effect of a traffic hotspot on the traffic of neighboring nodes and global performance degradation is investigated. To analyze the per-node traffic, some per-node traffic metrics are introduced and one of them is selected for the rest of work. In an effort to gain deep understanding of the issue of traffic analysis of faulty networks, this paper is the first attempt to investigate per-node traffic around fault regions.

show abstract

Section: Per-node Traffic Metricmentioning

confidence: 99%