2014 IEEE 34th International Conference on Distributed Computing Systems 2014
DOI: 10.1109/icdcs.2014.32
|View full text |Cite
|
Sign up to set email alerts
|

Scalable Traffic-Aware Virtual Machine Management for Cloud Data Centers

Abstract: Abstract-Virtual Machine (VM) management is a powerful mechanism for providing elastic services over Cloud Data Centers (DC)s. At the same time, the resulting network congestion has been repeatedly reported as the main bottleneck in DCs, even when the overall resource utilization of the infrastructure remains low. However, most current VM management strategies are traffic-agnostic, while the few that are traffic-aware only concern a static initial allocation, ignore bandwidth oversubscription, or do not scale.… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

1
24
0
4

Year Published

2015
2015
2019
2019

Publication Types

Select...
6
1

Relationship

2
5

Authors

Journals

citations
Cited by 34 publications
(29 citation statements)
references
References 21 publications
1
24
0
4
Order By: Relevance
“…(3) is satisfied, i.e., given the locally observed traffic, a VM u individually tests the candidate servers (for new placement) and migrates only when the benefit outweighs the migration cost c m . We refer interested readers to [16] in which we have formulated and proved the S-CORE scheme.…”
Section: A S-core Algorithmmentioning
confidence: 99%
See 2 more Smart Citations
“…(3) is satisfied, i.e., given the locally observed traffic, a VM u individually tests the candidate servers (for new placement) and migrates only when the benefit outweighs the migration cost c m . We refer interested readers to [16] in which we have formulated and proved the S-CORE scheme.…”
Section: A S-core Algorithmmentioning
confidence: 99%
“…= 0 do 12: newLocation ← GETDESTLOC(f low) 13: newT otalCost ← 0 14: for all f lows do 15: bytes ← GETFLOWBYTES(f low) 16: dest ← GETDEST(f low) 17: weight ← GETWEIGHT(newLocation, dest) 18: commCost ← bytes × weight 19: newT otalCost ← newT otalCost + commCost 20: end for 21: if newT otalCost < totalCost then 22: return newLocation ⊲ migrate! 23: end if 24: f low, cost ← GETHIGHESTCOMMFLOW(f lows) 25: end while hypervisors to send and receive tokens [16]. Instead, SDN handles dynamic environments too, as it monitors and reacts to real-time changes and automatically updates the relevant network parameters.…”
Section: B Sdn Dependenciesmentioning
confidence: 99%
See 1 more Smart Citation
“…In order to reduce congestion in the core layers of DC network, effective VM management schemes cluster VMs to confine traffic in lower layers of the network such that as much traffic as possible is routed only over the edge layer (which is not oversubscribed) [15] [12]. As a result, VMs as well as middleboxes, which communicate and exchange packets more often and intensively, are collocated in order to keep traffic within the edge layer boundaries.…”
Section: Migration Rulementioning
confidence: 99%
“…We can use some existing centralized algorithms to approximately maximize the total gained utility by migration, e.g., [24], [25]. However, the computation times of those algorithms are unacceptable for DCs, especially considering the large scales of servers, VMs, switches and millions of traffic flows [12]. In this section, we design a decentralized heuristic scheme to perform policyaware VMs migration.…”
Section: Plan Algorithmsmentioning
confidence: 99%