Authors:
Anna Schwanengel
;
Michael C. Jaeger
and
Uwe Hohenstein
Affiliation:
Siemens AG, Germany
Keyword(s):
Monitoring of Services, Quality of Service, Service Level Agreements, Load Balancing, Cloud Architecture.
Related
Ontology
Subjects/Areas/Topics:
Cloud Computing
;
Cloud Computing Architecture
;
Cloud Computing Enabling Technology
;
Energy and Economy
;
Fundamentals
;
Load Balancing in Smart Grids
;
Monitoring of Services, Quality of Service, Service Level Agreements
;
Smart Grids
Abstract:
Nowadays, cloud computing promises to supply a theoretically infinite resource amount, while enabling instance elasticity. However, using extra capacity requires organizational activities and leads to costs. To keep this overhead minimal, adding and releasing resources need to be well-scheduled. Therefore, it is inevitable to prepare an appropriate allocation automatism that considers differences in dynamics and price models of cloud computing. We present the idea for a system that effectively manages varying loads with regard to emerging costs, provisioning time, and customer service level agreements (SLAs). Contrary to existing, threshold-based solutions, our approach considers system observations oft he past, domain-specific load behaviours as well as external knowledge. That way, the system detects load patterns and adapts accordingly.