Multi-level Elasticity Control of Cloud Services Georgiana Copil, Daniel Moldovan, Hong-Linh Truong, Schahram Dustdar Distributed Systems Group, Vienna University of Technology
Multi-level Elasticity Control of Cloud Services -- ICSOC 2013
Dec 04, 2014
Presentation given at ICSOC 2013
Abstract: Fine-grained elasticity control of cloud services has to deal with multiple elasticity perspectives (quality, cost, and resources). We propose a cloud services elasticity control mechanism that considers the service structure for controlling the cloud service elasticity at multiple levels, by firstly defining an abstract composition model for cloud services and enabling multi-level elasticity control. Secondly, we define mechanisms for solving conflicting elasticity requirements and generating action plans for elasticity control. Using the defined concepts and mechanisms we develop a runtime system supporting multiple levels of elasticity control and validate the resulted prototype through experiments.
Abstract: Fine-grained elasticity control of cloud services has to deal with multiple elasticity perspectives (quality, cost, and resources). We propose a cloud services elasticity control mechanism that considers the service structure for controlling the cloud service elasticity at multiple levels, by firstly defining an abstract composition model for cloud services and enabling multi-level elasticity control. Secondly, we define mechanisms for solving conflicting elasticity requirements and generating action plans for elasticity control. Using the defined concepts and mechanisms we develop a runtime system supporting multiple levels of elasticity control and validate the resulted prototype through experiments.
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Multi-level Elasticity Control of Cloud Services
Georgiana Copil, Daniel Moldovan, Hong-Linh Truong, Schahram Dustdar
Distributed Systems Group, Vienna University of Technology
ICSOC, April 9, 2023 2
Overview
Motivation
Mapping Services Structures to Elasticity Metrics
Multi-level Control Runtime
Experiments
Conclusions and Future Work
ICSOC, April 9, 2023 3
Motivation
Traditional approach to cloud service control– Consider specific types of cloud services– Assume optimization strategies on behalf of the user– Do not consider cloud service structure
Tiramola [1] Control of NoSQL Clusters KingFisher [2] Cost-aware Provisioning Cloud Applications Auto-Scaling [3]
ICSOC, April 9, 2023 4
Our Approach
Use multi-level elasticity requirements, for knowing how to control the cloud service
Place modeling the cloud service and its environment at the center of the approach
Generate plans of abstract actions for elasticity control
ICSOC, April 9, 2023 5
Our Approach
Use multi-level elasticity requirements, for knowing how to control the cloud service
Place modeling the cloud service and its environment at the center of the approach
Generate plans of abstract actions for elasticity control
ICSOC, April 9, 2023 6
High Level Description of Elasticity Requirements
SYBL (Simple Yet Beautiful Language) for specifying elasticity requirements
SYBL-supported requirement levels– Cloud Service Level– Service Topology Level– Service Unit Level– Relationship Level– Programming/Code Level
#SYBL.CloudServiceLevelCons1: CONSTRAINT responseTime < 5 ms Cons2: CONSTRAINT responseTime < 10 ms WHEN nbOfUsers > 10000Str1: STRATEGY CASE fulfilled(Cons1) OR fulfilled(Cons2): minimize(cost)
#SYBL.ServiceUnitLevelStr2: STRATEGY CASE ioCost < 3 Euro : maximize( dataFreshness )
#SYBL.CodeRegionLevelCons4: CONSTRAINT dataAccuracy>90% AND cost<4 Euro
[Georgiana Copil, Daniel Moldovan, Hong-Linh Truong, Schahram Dustdar, "SYBL: an Extensible Language for Controlling Elasticity in Cloud Applications", 13th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid), May 14-16, 2013, Delft, Netherlands]
ICSOC, April 9, 2023 7
Our Approach
Use multi-level elasticity requirements, for knowing how to control the cloud service
Place modeling the cloud service and its environment at the center of the approach
Generate plans of abstract actions for elasticity control
ICSOC, April 9, 2023 8
Mapping Services Structures to Elasticity Metrics
ICSOC, April 9, 2023 9
Our Approach
Use multi-level elasticity requirements, for knowing how to control the cloud service
Place modeling the cloud service and its environment at the center of the approach
Generate plans of abstract actions for elasticity control
ICSOC, April 9, 2023 10
Multi-level Control Runtime:Generating Elasticity Control Plans
Cloud Providers/Tools must support higher and richer APIs for elasticity controls
ICSOC, April 9, 2023 11
Multi-level Control Runtime: Elasticity Control Prototype rSYBL
ICSOC, April 9, 2023 12
Experiments - Setup
Test Infrastructure:– Local cloud running OpenStack– Ganglia and Hyperic SIGAR for monitoring– JClouds for controlling virtual machine instances.
ICSOC, April 9, 2023 13
Experiments – Results [1/1]
Configuration
Controllers
DBNodes
Total execution time
Cost
Config1 1 3 578.4 s 0.48
Config2 1 6 472.1 s 0.91
Config3 2 2 382.4 s 0.42
Config4 3 7 372.2 s 0.72
Service unit level
Service topologylevel
ICSOC, April 9, 2023 14
Experiments – Results [1/1]
Configuration
Controllers
DBNodes
Total execution time
Cost
Config1 1 3 578.4 s 0.48
Config2 1 6 472.1 s 0.91
Config3 2 2 382.4 s 0.42
Config4 3 7 372.2 s 0.72
Service unit level
Service topologylevel
Configuration
Controllers
DBNodes
Workload Total execution time
Cost
Config1 1 3 Workload 1 44 min 2.92
Config3 2 2 Workload 1 28.4 min 1.88
Config1 1 3 Workload 2 >3h+errors >12
Config3 2 2 Workload 2 102.75 min 6.88
Service unit level
Service topologylevel
ICSOC, April 9, 2023 15
Experiments – Results [2/2]
ICSOC, April 9, 2023 16
SYBL & MELA
ICSOC, April 9, 2023 17
Conclusion and Future Work
Using SYBL, cloud providers could sell elasticity as a service to cloud consumers
SYBL and its runtime rSYBL enable multi-level elasticity control of cloud services
Future work– Elasticity behavior analysis– New/improved algorithms for the decision process
Visit SYBL webpage– http://www.infosys.tuwien.ac.at/research/viecom/SYBL– Tomorrow demo session: SYBL+MELA Demo
ICSOC, April 9, 2023 18
Thanks for your attention!
Georgiana Copil
[email protected]://www.infosys.tuwien.ac.at/staff/ecopil/
Distributed Systems GroupVienna University of TechnologyAustria
ICSOC, April 9, 2023 19
References
1. Dimitrios Tsoumakos, Ioannis Konstantinou, Christina Boumpouka, Spyros Sioutas, Nectarios Koziris, "Automated, Elastic Resource Provisioning for NoSQL Clusters Using TIRAMOLA," CCGRID, pp.34-41, 2013 13th IEEE/ACM International Symposium on Cluster, Cloud, and Grid Computing, 2013
2. Upendra Sharma; Shenoy, P.; Sahu, S.; Shaikh, A., "A Cost-Aware Elasticity Provisioning System for the Cloud," Distributed Computing Systems (ICDCS), 2011 31st International Conference on , vol., no., pp.559,570, 20-24 June 2011, doi: 10.1109/ICDCS.2011.59
3. Jing Jiang; Jie Lu; Guangquan Zhang; Guodong Long, "Optimal Cloud Resource Auto-Scaling for Web Applications," Cluster, Cloud and Grid Computing (CCGrid), 2013 13th IEEE/ACM International Symposium on , vol., no., pp.58,65, 13-16 May 2013doi: 10.1109/CCGrid.2013.73
Related Documents
![ShuttleDB: Database-Aware Elasticity in the Cloud · cloud elasticity mechanisms such as VM migration or VM replication (e.g., Amazon’s auto-scaling [2]), the approach does not](https://static.cupdf.com/doc/110x72/5f0b23ee7e708231d42f0c4f/shuttledb-database-aware-elasticity-in-the-cloud-cloud-elasticity-mechanisms-such.jpg)
