Extending OpenStack Monasca for Predictive Elasticity Control
Traditional auto-scaling approaches are conceived as reactive automations, typically triggered when predefined thresholds are breached by resource consumption metrics. Managing such rules at scale is cumbersome, especially when resources require non-negligible time to be instantiated. This paper int...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Tsinghua University Press
2024-06-01
|
| Series: | Big Data Mining and Analytics |
| Subjects: | |
| Online Access: | https://www.sciopen.com/article/10.26599/BDMA.2023.9020014 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832573274529202176 |
|---|---|
| author | Giacomo Lanciano Filippo Galli Tommaso Cucinotta Davide Bacciu Andrea Passarella |
| author_facet | Giacomo Lanciano Filippo Galli Tommaso Cucinotta Davide Bacciu Andrea Passarella |
| author_sort | Giacomo Lanciano |
| collection | DOAJ |
| description | Traditional auto-scaling approaches are conceived as reactive automations, typically triggered when predefined thresholds are breached by resource consumption metrics. Managing such rules at scale is cumbersome, especially when resources require non-negligible time to be instantiated. This paper introduces an architecture for predictive cloud operations, which enables orchestrators to apply time-series forecasting techniques to estimate the evolution of relevant metrics and take decisions based on the predicted state of the system. In this way, they can anticipate load peaks and trigger appropriate scaling actions in advance, such that new resources are available when needed. The proposed architecture is implemented in OpenStack, extending the monitoring capabilities of Monasca by injecting short-term forecasts of standard metrics. We use our architecture to implement predictive scaling policies leveraging on linear regression, autoregressive integrated moving average, feed-forward, and recurrent neural networks (RNN). Then, we evaluate their performance on a synthetic workload, comparing them to those of a traditional policy. To assess the ability of the different models to generalize to unseen patterns, we also evaluate them on traces from a real content delivery network (CDN) workload. In particular, the RNN model exhibites the best overall performance in terms of prediction error, observed client-side response latency, and forecasting overhead. The implementation of our architecture is open-source. |
| format | Article |
| id | doaj-art-28881375fbab4336baca216483eae923 |
| institution | Kabale University |
| issn | 2096-0654 |
| language | English |
| publishDate | 2024-06-01 |
| publisher | Tsinghua University Press |
| record_format | Article |
| series | Big Data Mining and Analytics |
| spelling | doaj-art-28881375fbab4336baca216483eae9232025-02-02T04:59:12ZengTsinghua University PressBig Data Mining and Analytics2096-06542024-06-017231533910.26599/BDMA.2023.9020014Extending OpenStack Monasca for Predictive Elasticity ControlGiacomo Lanciano0Filippo Galli1Tommaso Cucinotta2Davide Bacciu3Andrea Passarella4Scuola Normale Superiore, Pisa 56126, ItalyScuola Normale Superiore, Pisa 56126, Italythe Real-Time Systems Laboratory (RETIS), Telecommunications, Computer Engineering, and Photonics Institute (TeCIP), Scuola Superiore Sant’Anna, Pisa 56127, ItalyDepartment of Computer Science, University of Pisa, Pisa 56127, ItalyNational Research Council, Pisa56127, ItalyTraditional auto-scaling approaches are conceived as reactive automations, typically triggered when predefined thresholds are breached by resource consumption metrics. Managing such rules at scale is cumbersome, especially when resources require non-negligible time to be instantiated. This paper introduces an architecture for predictive cloud operations, which enables orchestrators to apply time-series forecasting techniques to estimate the evolution of relevant metrics and take decisions based on the predicted state of the system. In this way, they can anticipate load peaks and trigger appropriate scaling actions in advance, such that new resources are available when needed. The proposed architecture is implemented in OpenStack, extending the monitoring capabilities of Monasca by injecting short-term forecasts of standard metrics. We use our architecture to implement predictive scaling policies leveraging on linear regression, autoregressive integrated moving average, feed-forward, and recurrent neural networks (RNN). Then, we evaluate their performance on a synthetic workload, comparing them to those of a traditional policy. To assess the ability of the different models to generalize to unseen patterns, we also evaluate them on traces from a real content delivery network (CDN) workload. In particular, the RNN model exhibites the best overall performance in terms of prediction error, observed client-side response latency, and forecasting overhead. The implementation of our architecture is open-source.https://www.sciopen.com/article/10.26599/BDMA.2023.9020014elasticity controlauto-scalingpredictive operationsmonitoringopenstackmonasca |
| spellingShingle | Giacomo Lanciano Filippo Galli Tommaso Cucinotta Davide Bacciu Andrea Passarella Extending OpenStack Monasca for Predictive Elasticity Control Big Data Mining and Analytics elasticity control auto-scaling predictive operations monitoring openstack monasca |
| title | Extending OpenStack Monasca for Predictive Elasticity Control |
| title_full | Extending OpenStack Monasca for Predictive Elasticity Control |
| title_fullStr | Extending OpenStack Monasca for Predictive Elasticity Control |
| title_full_unstemmed | Extending OpenStack Monasca for Predictive Elasticity Control |
| title_short | Extending OpenStack Monasca for Predictive Elasticity Control |
| title_sort | extending openstack monasca for predictive elasticity control |
| topic | elasticity control auto-scaling predictive operations monitoring openstack monasca |
| url | https://www.sciopen.com/article/10.26599/BDMA.2023.9020014 |
| work_keys_str_mv | AT giacomolanciano extendingopenstackmonascaforpredictiveelasticitycontrol AT filippogalli extendingopenstackmonascaforpredictiveelasticitycontrol AT tommasocucinotta extendingopenstackmonascaforpredictiveelasticitycontrol AT davidebacciu extendingopenstackmonascaforpredictiveelasticitycontrol AT andreapassarella extendingopenstackmonascaforpredictiveelasticitycontrol |