Abstract
Distributed hierarchical network monitoring model has been proposed to solve scalability problem of centralized model. In this distributed model, a top-level monitoring manager, called main manager, obtains aggregate management information from mid-level managers, named domain managers, forming a hierarchical structure. However, if some of monitoring managers crash, network elements cannot be continuously and correctly monitored until the managers are repaired. To address this important, but previously unresolved issue, this paper presents a new fault-tolerance protocol for domain managers, named DMFTP, allowing the managers to efficiently utilize their organization structure. Therefore, this protocol can minimize failure detection overhead and the number of live managers affected by each manager node crash. Also, it tolerates concurrent manager failures and, after the failed managers have been repaired, ensures their immediate and consistent recovery.
| Original language | English |
|---|---|
| Pages (from-to) | 504-513 |
| Number of pages | 10 |
| Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Volume | 2659 |
| Publication status | Published - 2003 Dec 1 |
Fingerprint
ASJC Scopus subject areas
- Computer Science(all)
- Biochemistry, Genetics and Molecular Biology(all)
- Theoretical Computer Science
Cite this
Low-cost fault-tolerance protocol for large-scale network monitoring. / Ahn, JinHo; Min, Sung-Gi; Choi, YoungIl; Lee, ByungSun.
In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 2659, 01.12.2003, p. 504-513.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Low-cost fault-tolerance protocol for large-scale network monitoring
AU - Ahn, JinHo
AU - Min, Sung-Gi
AU - Choi, YoungIl
AU - Lee, ByungSun
PY - 2003/12/1
Y1 - 2003/12/1
N2 - Distributed hierarchical network monitoring model has been proposed to solve scalability problem of centralized model. In this distributed model, a top-level monitoring manager, called main manager, obtains aggregate management information from mid-level managers, named domain managers, forming a hierarchical structure. However, if some of monitoring managers crash, network elements cannot be continuously and correctly monitored until the managers are repaired. To address this important, but previously unresolved issue, this paper presents a new fault-tolerance protocol for domain managers, named DMFTP, allowing the managers to efficiently utilize their organization structure. Therefore, this protocol can minimize failure detection overhead and the number of live managers affected by each manager node crash. Also, it tolerates concurrent manager failures and, after the failed managers have been repaired, ensures their immediate and consistent recovery.
AB - Distributed hierarchical network monitoring model has been proposed to solve scalability problem of centralized model. In this distributed model, a top-level monitoring manager, called main manager, obtains aggregate management information from mid-level managers, named domain managers, forming a hierarchical structure. However, if some of monitoring managers crash, network elements cannot be continuously and correctly monitored until the managers are repaired. To address this important, but previously unresolved issue, this paper presents a new fault-tolerance protocol for domain managers, named DMFTP, allowing the managers to efficiently utilize their organization structure. Therefore, this protocol can minimize failure detection overhead and the number of live managers affected by each manager node crash. Also, it tolerates concurrent manager failures and, after the failed managers have been repaired, ensures their immediate and consistent recovery.
UR - http://www.scopus.com/inward/record.url?scp=35248877563&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=35248877563&partnerID=8YFLogxK
M3 - Article
AN - SCOPUS:35248877563
VL - 2659
SP - 504
EP - 513
JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SN - 0302-9743
ER -