2013 13th IEEE/ACM International Symposium on Cluster, Cloud, and Grid Computing 2013
DOI: 10.1109/ccgrid.2013.74
|View full text |Cite
|
Sign up to set email alerts
|

Autoflex: Service Agnostic Auto-scaling Framework for IaaS Deployment Models

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
5
0
3

Year Published

2016
2016
2023
2023

Publication Types

Select...
4
3
1

Relationship

0
8

Authors

Journals

citations
Cited by 28 publications
(8 citation statements)
references
References 11 publications
0
5
0
3
Order By: Relevance
“…This strategy is commonly used by auto-scalers that only support vertical scaling, as for a single machine, resource usage can substitute workload intensity. Some proposals [Islam et al 2010;Caron et al 2011;Almeida Morais et al 2013] that target horizontal scaling also follow this strategy to accomplish both workload prediction and resource estimation together. [Gong et al 2010] used signal processing to discover the longest repeating pattern of resource usage and then relied on dynamic time warping (DTW) algorithm to make the prediction.…”
Section: Proactive Scalingmentioning
confidence: 99%
See 1 more Smart Citation
“…This strategy is commonly used by auto-scalers that only support vertical scaling, as for a single machine, resource usage can substitute workload intensity. Some proposals [Islam et al 2010;Caron et al 2011;Almeida Morais et al 2013] that target horizontal scaling also follow this strategy to accomplish both workload prediction and resource estimation together. [Gong et al 2010] used signal processing to discover the longest repeating pattern of resource usage and then relied on dynamic time warping (DTW) algorithm to make the prediction.…”
Section: Proactive Scalingmentioning
confidence: 99%
“…[Yazdanov and Fetzer 2012] utilized an auto-regressive (AR) method to predict short-term CPU usage. [Almeida Morais et al 2013] employed multiple time series algorithms to predict CPU usage, and based on their runtime accuracy, the best is selected. [Loff and Garcia 2014] also used various prediction algorithms.…”
Section: Proactive Scalingmentioning
confidence: 99%
“…The RL is realized by using parallel learning; that is, the authors intend to speed up agent's learning process of approximated model by learning in parallel, without visiting every state-action pair in a given environment. The approaches that rely on demand prediction (e.g., the Autoflex [25] and PRESS [80,129,42,60,90]) are also regarded as implicit search. This is because the autoscaling decision is directly predicted by the demand models, without the needs of reasoning and optimization.…”
Section: Search-based Optimizationmentioning
confidence: 99%
“…Ωστόσο, οι καθαρά Προληπτικές τεχνικές τείνουν να υστερούν στην αντιμετώπιση ξαφνικών φαινομένων στη διακύμανση του φόρτου. Για να ξεπεράσουν αυτή τη δυσκολία, κάποιες εργασίες [78,25,35] υιοθετούν μία υβριδική προσέγγιση. Οι εργασίες [24,70] προτείνουν τη χρήση μίας Αντιδραστικής προσέγγισης κατά την επέκταση προς τα επάνω και μίας Προληπτικής προσέγγισης κατά τη συρρίκνωση.…”
Section: ενεργοποίηση διαδικασίας λήψης αποφάσεωνunclassified
“…Οι εργασίες [25,24,94,95,39] χρησιμοποιούν προσομοίωση για το στάδιο της αξιολόγησης, όπως π.χ. χρήση της σουίτας στατιστικών R [18] ή του πλαισίου προσομοίωσης OMNeT++ [120].…”
Section: μεθοδολογία αξιολόγησηςunclassified