2017 IEEE 37th International Conference on Distributed Computing Systems (ICDCS) 2017
DOI: 10.1109/icdcs.2017.253
|View full text |Cite
|
Sign up to set email alerts
|

A Preventive Auto-Parallelization Approach for Elastic Stream Processing

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
11
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
5
1

Relationship

1
5

Authors

Journals

citations
Cited by 20 publications
(11 citation statements)
references
References 11 publications
0
11
0
Order By: Relevance
“…Kombi et al [79] published the AUTOSCALE approach that centrally adapts the parallelization degrees of the operators in an operator graph based on predictive input values. eir goal is to proactively avoid congestion within the operator graph but be resource optimal and avoid too frequent recon gurations.…”
Section: Centralized Elasticitymentioning
confidence: 99%
“…Kombi et al [79] published the AUTOSCALE approach that centrally adapts the parallelization degrees of the operators in an operator graph based on predictive input values. eir goal is to proactively avoid congestion within the operator graph but be resource optimal and avoid too frequent recon gurations.…”
Section: Centralized Elasticitymentioning
confidence: 99%
“…In addition, this website includes a comparison between Autoscale and Autoscale+. It gives an overview of the gap between the performance of the auto-parallelization strategy presented in [24] and results presented below. In the remainder of this section, the results described correspond to average values over 5 iterations for each configuration, thus lessening the the impact of punctual anomalies during tests.…”
Section: Experiments and Resultsmentioning
confidence: 99%
“…In [24] we defined a proactive approach, named Autoscale, to manage dynamically and automatically the parallelism degree of operators using indicators monitored on streams and operators. Our proposed algorithm decides which operators have to be reconfigured (scale-out or scale-in) and what their new parallelism degrees are.…”
Section: Parallelism Management With Autoscale+mentioning
confidence: 99%
See 2 more Smart Citations